Citation
Analysis of the relationship among personal factors, risk perception, and behavioral responses related to climate change in South Korea

Material Information

Title:
Analysis of the relationship among personal factors, risk perception, and behavioral responses related to climate change in South Korea
Creator:
Jung, Youn Suk
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English

Thesis/Dissertation Information

Degree:
Doctorate ( Doctor of philosophy)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
School of Public Affairs, CU Denver
Degree Disciplines:
Public affairs
Committee Chair:
Weible, Christopher
Committee Members:
Heikkila, Tanya
Swann, William
Lee, Kihoon

Notes

Abstract:
Despite global collaboration, climate change is getting worse. Fortunately, individuals’ daily actions, along with governments’ policies, can be alternatives for mitigating Greenhouse Gas (GHG) emissions and alleviating the negative impacts of climate change. However, empirical evidence about the process of individuals’ behavioral responses to address climate change, particularly the behavioral responses of the general public, is still insufficient. According to previous studies, climate change risk perception is deeply associated with individuals’ behavioral responses against climate change. To strengthen empirical evidence of how behavioral responses to cope with climate change within the general public are shaped, this dissertation investigates the causal relationships among personal factors, risk perception, and behavioral responses related to climate change. It hypothesizes that personal factors associated with climate change (i.e., experiences, knowledge, and policy assessment) affect climate change risk perception. Subsequently, higher risk perception is likely to intensify individuals’ behavioral responses to climate change. In addition, some personal factor variables (e.g., extreme weather event experiences and policy-related knowledge) may have a direct impact on individuals’ behavioral responses. To assess these hypotheses, both quantitative analyses at the national and local levels and qualitative analyses at the expert level are conducted in the South Korean context. South Korea is contributing to increasing GHG emissions as a result of economic growth over the past decades. First, for the national-level analysis, the measure of behavioral responses focuses on individuals’ GHG mitigation behaviors (e.g., saving energy and utilizing renewable energy). Data are collected from a web-based survey implemented in South Korea (2016, N = 1,500). The results of regression analysis, specifically Structural Equation Modeling (SEM), are largely consistent with this study’s expectations. Most personal factors, such as long-term climate change experiences, awareness of climate change, and the assessment of drought policy, significantly influence climate change risk perception. Subsequently, higher climate change risk perception, along with higher familiarity with a government GHG reduction plan, enhances behavioral responses to mitigate climate change. Meanwhile, extreme weather event experiences (e.g., flood and drought) do not have an impact on either risk perception or behavioral responses. Second, the local-level analysis focuses on individuals’ adaptation behaviors against drought as the measure of behavioral responses. Based on survey data derived from South Chungcheong Province, South Korea (2015, N = 515), the SEM regression analysis presents mixed results. Several personal factors, such as exposure to water saving information and the assessment of water infrastructure policy, are significantly correlated with drought risk perception. However, higher drought risk perception, unlike higher familiarity with a subsidy policy associated with water saving, fails to intensify behavioral responses to adapt to drought. Also, water problem experiences (e.g., limited water supply) have no relationship with both risk perception and behavioral responses. Third, the last empirical analysis utilizes expert knowledge from face-to-face interviews (2018, N = 13) and an online survey (2018, N = 98) with climate change experts. Their insights offer support for most of the study’s hypotheses. Regarding the gaps between this study’s hypotheses and the results of the quantitative analyses, the experts recognize that the significant relationship between experiences and risk perception requires several conditions, such as experiences with direct and significant damage, recent experiences, and instances with no possible solution for minimizing damage. Also, many of them are suspicious about the relationship between drought risk perception and behavioral responses against drought. This is because prioritized residential water supply policy and well-established response infrastructure generally ensure that individuals are not exposed to direct and severe damage triggered by drought, which makes individuals reluctant to take action. Given the context of South Korea, such as a high level of climate change adaptation capacity, the recent downward trend of damage brought by floods, and government’s top priority on residential water during drought, the identified gaps between the study’s findings and the hypotheses seem understandable. Overall, while risk perception plays an important role in enhancing behavioral responses to mitigate climate change, the precise role of risk perception in climate change adaptation may depend on the context of each country (e.g., the capacity of climate change adaptation and recent trends in extreme weather events). Despite some limitations, this study successfully offers empirical evidence of the theorized relationships among personal factors, risk perception, and behavioral responses associated with climate change, which can be important to understand when addressing climate change. The findings of this dissertation and mixed methods research not only contribute to advancing knowledge and the research methodology of the process of individuals’ behavioral responses to cope with climate change, but they also inform policymakers of critical clues as to how the general public voluntarily takes action to address climate change.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
Copyright [name of copyright holder or Creator or Publisher as appropriate]. Permission granted to University of Colorado Denver to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.

Downloads

This item has the following downloads:


Full Text
ANALYSIS OF THE RELATIONSHIP AMONG PERSONAL FACTORS, RISK
PERCEPTION, AND BEHAVIORAL RESPONSES RELATED TO CLIMATE CHANGE IN
SOUTH KOREA by
YOUN SUK JUNG
B.Ec., Chungnam National University, 2001 M.Ec., Chungnam National University, 2014
A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Doctor of Philosophy Public Affairs Program
2019


©2019
YOUN SUK JUNG ALL RIGHTS RESERVED
11


This thesis for the Doctor of Philosophy degree by Youn Suk Jung has been approved for the Public Affairs Program by
Christopher Weible, Chair Tanya Heikkila, Advisor William Swann Kihoon Lee
Date: July 27, 2019


Jung, Youn Suk (PhD, Public Affairs Program)
Analysis of the Relationship among Personal Factors, Risk Perception, and Behavioral Responses related to Climate Change in South Korea Thesis directed by Professor Tanya Heikkila
ABSTRACT
Despite global collaboration, climate change is getting worse. Fortunately, individuals’ daily actions, along with governments’ policies, can be alternatives for mitigating Greenhouse Gas (GHG) emissions and alleviating the negative impacts of climate change. However, empirical evidence about the process of individuals’ behavioral responses to address climate change, particularly the behavioral responses of the general public, is still insufficient.
According to previous studies, climate change risk perception is deeply associated with individuals’ behavioral responses against climate change. To strengthen empirical evidence of how behavioral responses to cope with climate change within the general public are shaped, this dissertation investigates the causal relationships among personal factors, risk perception, and behavioral responses related to climate change. It hypothesizes that personal factors associated with climate change (i.e., experiences, knowledge, and policy assessment) affect climate change risk perception. Subsequently, higher risk perception is likely to intensify individuals’ behavioral responses to climate change. In addition, some personal factor variables (e.g., extreme weather event experiences and policy-related knowledge) may have a direct impact on individuals’ behavioral responses. To assess these hypotheses, both quantitative analyses at the national and local levels and qualitative analyses at the expert level are conducted in the South Korean context. South Korea is contributing to increasing GHG emissions as a result of economic growth over the past decades.
IV


First, for the national-level analysis, the measure of behavioral responses focuses on individuals’ GHG mitigation behaviors (e.g., saving energy and utilizing renewable energy).
Data are collected from a web-based survey implemented in South Korea (2016, N = 1,500). The results of regression analysis, specifically Structural Equation Modeling (SEM), are largely consistent with this study’s expectations. Most personal factors, such as long-term climate change experiences, awareness of climate change, and the assessment of drought policy, significantly influence climate change risk perception. Subsequently, higher climate change risk perception, along with higher familiarity with a government GHG reduction plan, enhances behavioral responses to mitigate climate change. Meanwhile, extreme weather event experiences (e.g., flood and drought) do not have an impact on either risk perception or behavioral responses.
Second, the local-level analysis focuses on individuals’ adaptation behaviors against drought as the measure of behavioral responses. Based on survey data derived from South Chungcheong Province, South Korea (2015, N = 515), the SEM regression analysis presents mixed results. Several personal factors, such as exposure to water saving information and the assessment of water infrastructure policy, are significantly correlated with drought risk perception. However, higher drought risk perception, unlike higher familiarity with a subsidy policy associated with water saving, fails to intensify behavioral responses to adapt to drought. Also, water problem experiences (e.g., limited water supply) have no relationship with both risk perception and behavioral responses.
Third, the last empirical analysis utilizes expert knowledge from face-to-face interviews (2018, N = 13) and an online survey (2018, N = 98) with climate change experts. Their insights offer support for most of the study’s hypotheses. Regarding the gaps between this study’s hypotheses and the results of the quantitative analyses, the experts recognize that the significant
v


relationship between experiences and risk perception requires several conditions, such as experiences with direct and significant damage, recent experiences, and instances with no possible solution for minimizing damage. Also, many of them are suspicious about the relationship between drought risk perception and behavioral responses against drought. This is because prioritized residential water supply policy and well-established response infrastructure generally ensure that individuals are not exposed to direct and severe damage triggered by drought, which makes individuals reluctant to take action. Given the context of South Korea, such as a high level of climate change adaptation capacity, the recent downward trend of damage brought by floods, and government’s top priority on residential water during drought, the identified gaps between the study’s findings and the hypotheses seem understandable.
Overall, while risk perception plays an important role in enhancing behavioral responses to mitigate climate change, the precise role of risk perception in climate change adaptation may depend on the context of each country (e.g., the capacity of climate change adaptation and recent trends in extreme weather events). Despite some limitations, this study successfully offers empirical evidence of the theorized relationships among personal factors, risk perception, and behavioral responses associated with climate change, which can be important to understand when addressing climate change. The findings of this dissertation and mixed methods research not only contribute to advancing knowledge and the research methodology of the process of individuals’ behavioral responses to cope with climate change, but they also inform policymakers of critical clues as to how the general public voluntarily takes action to address climate change.
The form and content of this abstract are approved. I recommend its publication.
Approved: Tanya Heikkila
vi


TABLE OF CONTENTS
CHAPTER
I. INTRODUCTION..................................................................1
II. BACKGROUND...................................................................7
Fundamentals of Climate Change...............................................7
Beginning of Anthropogenic Climate Change................................7
Emergence of a Global Climate Change Organization.......................8
Global Efforts for Addressing Climate Change............................10
Conditions Surrounding Global Collaboration.............................13
Government and Individual’s Responses to Climate Change.....................19
Anthropogenic-Natural Hybrid Disaster...................................19
Difficulties in Government Responses to Climate Change..................21
Difficulties in Individual Responses to Climate Change..................23
Advantages of Individuals’ Behavioral Responses to Climate Change.......29
III. THEORY.....................................................................31
Risk Perception Driving Behavioral Responses................................31
Climate Change Risk.....................................................31
Role of Public Risk Perception..........................................33
Gaps in Climate Change Risk Perception Literature.......................36
Filling the Gaps in the Literature: Research Question...................38
Integrated Research Model...................................................40
Theoretical Foundation for Individuals’ Behavioral Responses............40
Research Modeling Approach..............................................45
vii


Hypotheses and Research Model..........................................47
Potential Issues in the Methodology....................................61
IV. NATIONAL-LEVEL ANALYSIS...................................................64
Research Contexts..........................................................64
Economic Development...................................................64
Greenhouse Gas Emissions...............................................65
Literatures related to Climate Change Risk Perception in South Korea...65
Research Methods...........................................................66
Data Collection........................................................66
Variable Description...................................................70
Data Analysis..........................................................74
Analysis Results...........................................................77
Socio-demographic Characteristics......................................77
Descriptive Results....................................................77
Bivariate Correlation Coefficients among Variables.....................79
Model Fit..............................................................81
Regression Analysis....................................................81
Alternative Model Specifications.......................................87
Policy Implications........................................................92
Limitations and Future Research Directions.................................95
Conclusion.................................................................97
V. LOCAL-LEVEL ANALYSIS......................................................99
viii
Research Contexts
99


Water Management and Climate Change.....................................99
Water Management in South Korea........................................100
Water Condition in South Chungcheong Province..........................101
Research Methods...........................................................103
Data Collection........................................................103
Variable Description...................................................105
Data Analysis..........................................................110
Analysis Results...........................................................Ill
Socio-demographic Characteristics......................................Ill
Descriptive Results....................................................112
Bivariate Correlation Coefficients among Variables.....................113
Model Fit..............................................................115
Regression Analysis....................................................115
Alternative Model Specifications.......................................122
Policy Implications........................................................123
Limitations and Future Research Directions.................................126
Conclusion.................................................................127
VI. EXPERT-LEVEL ANALYSIS.....................................................129
Research Contexts..........................................................129
Necessity of Mixed Methods Research....................................129
Advantages of Utilizing Experts........................................131
Core Components of Experts.............................................132
Expert Interviews..........................................................133
ix


Data Collection
133
Data Analysis........................................................135
Analysis Results.....................................................137
Expert Survey............................................................157
Data Collection......................................................157
Socio-demographic Characteristics....................................157
Analysis Results.....................................................158
Policy Implications......................................................162
Research Implications....................................................163
Conclusion...............................................................164
VII. CONCLUSION..............................................................167
Content of Chapter II—VI.................................................168
Attainment of Study Objectives...........................................172
Lessons for Policymakers.................................................174
Lessons for Researchers..................................................175
Epilogue.................................................................176
REFERENCES...................................................................177
APPENDIX.....................................................................212
A. Conceptual map for variables.........................................212
B. Key questions of the 2016 survey in South Korea......................214
C. Results of SEM regression analyses with different estimation methods (or options) in the national-level analysis.........................................................216
x


D. Direct, indirect, and total effects of independent and control variables on behavioral
responses to mitigate climate change....................................217
E. Relationship between personal factors and socio-demographic characteristics in the
national and local level analyses...........................................218
F. Key questions of the 2015 survey in north-west area of South Chungcheong
Province......................................................................220
G. Results of SEM regression analyses with different estimation methods (or options)
in the local-level analysis...................................................222
H. Direct, indirect, and total effects of independent and control variables on behavioral
responses to adapt to drought.................................................224
I Main questions of the climate change expert interview.........................225
J. Part of the codebook for interview transcription..............................226
xi


LIST OF TABLES
TABLE
1. Significant socio-demographic factors of previous studies..............................59
2. 2016 survey respondents’ socio-demographic characteristics............................69
3. Inter-correlations of key variables in the national-level analysis....................80
4. Results of SEM regression analysis about the relationship among personal factors, risk
perception, and behavioral responses related to climate change.........................83
5. Results of GSEM regression analysis about the relationship among personal factors, risk
perception, and each behavioral response associated with climate change................89
6. 2015 survey respondents’ socio-demographic characteristics............................104
7. Correlation matrix of key variables in the local-level analysis.......................114
8. Results of SEM regression analyses about the relationship among personal factors, risk
perception, and behavioral responses related to drought................................116
9. Summary of the interviewees’ responses about the relationship among personal factors, risk
perception, and behavioral responses in the mitigation and adaptation contexts.........140
10. Possible reasons for the insignificant relationship between experiences and risk perception 143
11. Summary of the survey to members of the KWRA..........................................159
12. Descriptions about the relationship between climate change experiences and climate change
risk perception.......................................................................160
13. Summary of three empirical analyses...................................................171
xii


LIST OF FIGURES
FIGURE
1. Kates’ (1971) general systems model of human adjustment to natural hazards...............41
2. Rogers’ (1983) protection motivation theory.............................................42
3. Grothmann and Patt’s (2005) model of private proactive adaptation to climate change.....44
4. Research model of factors affecting individuals’ behavioral responses to address climate
change.....................................................................................60
5. Pathways in the national-level research model...........................................76
6. Results of four long-term climate change experiences....................................78
7. Results of individuals’ behavioral responses to address climate change..................79
8. Significant pathways in the national-level research model...............................84
9. Pathways in the local-level research model..............................................Ill
10. Results of seven water saving efforts than usual.......................................113
11. Significant pathways in the daily action and device utilization models.................117
12. Amount of damage caused by typhoons and heavy rains in South Korea (From 2002 to 2015) 145
13. Changes in the relationship between short-term experiences and risk perception related to
climate change..........................................................................146
xiii


LIST OF ABBREVIATIONS
ADF: Asymptotically Distribution-Free CAT: Climate Action Tracker CFI: Comparative Fit Index COP: Conference of the Parties GHG: Green House Gas
GSEM: Generalized Structural Equation Modeling
ICT: Information and Communication Technologies
INDCs: Intended Nationally Determined Contributions
IPCC: Intergovernmental Panel on Climate Change
K-water: Korea Water Resources Corporation
KWRA: Korea Water Resources Association
MOLIT: Ministry of Infrastructure and Transport
NASA: National Aeronautics and Space Administration
OECD: Organization for Economic Cooperation and Development
PMT: Protection Motivation Theory
RCP: Representative Concentration Pathway
RMSEA: Root Mean Square Error of Approximation
SEM: Structural Equation Modeling
SRMR: Standardized Root Mean Squared Residual
TLI: Tucker-Lewis Index
UN: United Nations
UNFCCC: United Nations Framework Convention on Climate Change
xiv


CHAPTER I
INTRODUCTION
Climate change is not a new issue. As early as 1896, the Swedish chemist Svante Arrhenius explored the relationship between atmospheric carbon dioxide and air temperature to explain the existence of ice ages (Arrhenius, 1896). In 1964, social ecologist Murray Bookchin warned of the probability of anthropogenic climate change appearing and its potential adverse impacts (Bookchin, 1964). However, this issue failed to bring heightened attention from the general public and the international community for a long time. Over two decades later, the Intergovernmental Panel on Climate Change (IPCC), which is the international organization for providing a scientific basis related to climate change to governments at all levels, was set up in 1988 (IPCC, 2017). Since then, the IPCC has continually warned of the risks of climate change through the IPCC Assessment Reports (IPCC, 2014a). As a result, it has been highlighted as a major issue around the world.
Global collaboration to address climate change has remained sluggish due to complicated political (e.g., numerous stakeholders and the absence of legitimate and coercive global organizations), economic (e.g., enormous costs to mitigate and adapt to climate change), and social (e.g., each nation’s different level of public concern and vulnerability to climate change) circumstances. (Bodansky, 2012; Hagen, 2015; Klein, Schipper, & Dessai, 2005; Nyong, 2009; Stern, 2006). Also, technical advancements linked to the reduction of Greenhouse Gas (GHG) emissions and the use of renewable energy are still insufficient and have not provided momentum to global collaboration. In 2015, the international community seemed to bear fruitful results by establishing the first global action plan to curb climate change. This Paris Climate Agreement was endorsed by one hundred and ninety-five countries (European Commission,
1


2017), and the United States (U.S.) and China were viewed as leaders pushing this agreement’s fulfillment (Bodansky, 2016). A recent challenge to the agreement came early in 2017 when President Donald Trump declared that the U.S. would withdraw from the agreement (Ellis,
2017). The prospect of international climate policy was dampened by this event (Shear, 2017).
Even though climate change is regarded as a global level risk by both the scientific and non-scientific worlds, the national, local, and individual levels typically bear the costs of most impacts (Altschuler & Brownlee, 2016; Mendelsohn, Dinar, & Williams, 2006). To effectively alleviate the adverse effects of climate change, the world needs not only efforts at the global level, but also those at the national, local, and individual levels. However, it is very difficult for both national and local policymakers to design and carry out effective and efficient climate change policies due to several challenges, such as the uncertainty around its predicted trajectories and impacts (Stern, 2006; Tompkins & Adger, 2005), juggling mitigation and adaptation strategies (Biesbroek, Swart, & Van der Knaap, 2009; Kane & Yohe, 2000), and the relatively low priority of climate change among national issues (Dunlap & Saad, 2001; Leiserowitz, 2005). In particular, policymakers in countries that suffer from a lack of resources (e.g., funds, experts, and technologies) are facing much more serious conditions. Fortunately, many existing studies (e.g., Clayton et al., 2015; Semenza et al., 2008; Shove, 2010) advise that not only governments’ policies, but also individuals’ behavioral responses can be alternatives for alleviating the adverse impacts of climate change. Further, the more behavioral responses to address climate change that individuals take, the less of a burden policymakers will feel in shaping and implementing climate policies. Thus, revealing how to inspire individuals to more actively participate in coping actions against climate change might be no simple task, but be very meaningful. This dissertation investigates the process of individuals’ behavioral responses to address climate change.
2


Recently, many researchers have paid attention to the question, what motivates individuals to take behavioral responses to curb climate change? Most of them focus on personal factors (e.g., experience and knowledge) and climate change risk perception. It is true that public risk perception has been spotlighted by scholars in that it can inform policymakers of public concerns related to particular risks and the public’s preference for public policies (Pidgeon,
1998; Pidgeon & Gregory, 2004).1 At the same time, however, some previous studies (e.g., Battaglini et al., 2009; O’Connor, Bard, & Fisher, 1999; Spence et al., 2011) have noticed a different role of public risk perception, which is as a trigger of individuals ’ behavioral responses to address risks. They note that higher levels of perceived climate change risk can lead to changes in personal actions. Also, personal factors are expected to influence behavioral responses to cope with climate change via risk perception or directly.
Unfortunately, previous empirical research projects do not provide enough evidence about the process of individuals’ behavioral responses with respect to mitigating climate change and adapting to different types of climate change impacts (e.g., drought and flood risks). Most existing studies focus on the simple relationship between personal factors and climate change risk perception (e.g., Brody et al., 2008; Lowe et al., 2006; Slovic et al., 2004) or between climate change risk perception and behavioral responses to address climate change (e.g., Battaglini et al., 2009; Joshi et al., 2013; Roco et al., 2015). Although a few research projects have investigated the relationship among personal factors, climate change risk perception, and behavioral responses against climate change, they are usually based on specific industry contexts, such as agriculture (Niles, Lubell, & Haden, 2013), or local level data (e.g., flood and
1 Public (risk) perception has been studied in diverse fields (e.g., health, technology, and environment) because policy-making is affected by what the public perceives (Bostrom et al., 1994; Leiserowitz, 2006; Ruddell et al., 2012). In the climate change field, public risk perception can play a significant role in both shaping climate policies (Brody et al., 2008; Liu, Smith, & Safi, 2014) and generating public support for those policies (Leiserowitz, 2005).
3


drought-prone areas) (Whitmarsh, 2008). The findings of these studies are not necessarily applicable to behavioral responses of the general public at the national level. Also, while a number of studies (e.g., Lindell & Hwang, 2008; Takao etal., 2004) explored risk perception and self-protective behaviors by individuals against flood, research projects investigating personal risk perception and behavioral responses against drought are very rare (e.g., Deng, Wang, & Yousefpour, 2017). To fill the gap in the literature and support policymakers’ decisionmaking linked to climate policy, this dissertation tries to answer the following research question: How do personal factors (experiences, knowledge, and policy assessment) and risk perception related to climate change influence behavioral responses to address climate change? Obviously, the aim of this study is not to explain all of the potential factors that drive individuals’ behavioral responses to cope with climate change, but to shed light on this theoretical relationship and in doing so better articulate the role of risk perception in mediating the influence of individuals’ experience, knowledge, and policy assessments of climate change.
This question is answered based on the case of South Korea, which is experiencing the rapid increase in GHG emissions in return for economic growth over the past decades. More specifically, the answer is derived from three different contexts: a) climate change risk perception and personal behaviors to mitigate climate change among individuals at the national level; b) drought risk perception and personal behaviors to adapt to drought among individuals at the local level; and c) significant issues affecting the relationship among personal factors, risk perception, and behavioral responses related to climate change as perceived by experts. Prior to empirical analyses, five hypotheses, including the positive correlation between risk perception and behavioral responses related to climate change, are proposed based on literature reviews.
The works of Kates (1971), Rogers (1983), and Grothmann and Patt (2005) provide a theoretical
4


foundation for developing the hypotheses. Two data sets, provided by the Korea Water Resources Corporation (K-water), are utilized to conduct quantitative analyses at the national and local levels. K-water collected national level data by a Web-based survey (2016, N = 1,500) in South Korea, and collected local level data from a face-to-face survey (2015, N = 515) in north-west area of South Chungcheong Province, South Korea.2 Regression analysis, specifically structural equation modeling, is employed to evaluate the data. To reinforce the findings of the two cross-sectional analyses, this study conducts qualitative analyses based on an expert survey (2018, N = 98) and interviews (2018, N = 13) collected from the members of the Korea Water Resource Association, the K-water institute, and other academic organizations.3
By providing empirical evidence about the relationship among personal factors, risk perception, and behavioral responses to address climate change at the national and local levels along with expert knowledge, academically, this study contributes to improving knowledge of behavioral responses against climate change as well as advancing the methodology of unveiling that relationship by using mixed-methods research. In addition, it is meaningful in that this study will be an initial step toward revealing how climate change risk perception serves as a link between personal factors and behavioral responses related to climate change in South Korea. Practically, this dissertation informs policymakers of critical clues as to how the general public voluntarily takes action to address climate change. Policymakers both in South Korea and in countries with similar socio-economic conditions can benefit the most from this study’s findings.
2 Two survey companies, the opensurvey (‘www.opensurvey.co.kr’) and the Gallup Korea (‘www.gallup.co.kr’), conducted the Web-based survey in 2016 and the face-to-face survey in 2015 respectively, by signing survey contracts with K-water.
3 Those data include expert knowledge not only about the causal relationship among personal factors, risk perception, and behavioral responses to address climate change, but also about significant issues surrounding the relationship among them (e.g., obstacles which hinder the relationship and governmental policies to encourage individuals’ behavioral responses) within the general public.
5


This dissertation consists of seven chapters, including this introduction chapter.
Two literature reviews. The second chapter covers the background literature explaining the fundamentals of climate change and climate policy and individuals’ behavioral responses against climate change risk. It helps understand the obstacles of global collaboration to address climate change, policymakers’ difficulties in managing climate policy, and the features of individuals’ responses to cope with climate change. The third chapter presents existing theories and empirical studies dealing with the relationship among personal factors, risk perception, and behavioral responses related to climate change. A research question and hypotheses are provided in this chapter. Also, a research model based on the hypotheses is set up here.
Three empirical analyses. The fourth and fifth chapters quantitatively test the hypotheses at the national and local levels, respectively. While the former focuses on climate change risk perception and behavioral responses to mitigate climate change, the latter pays attention to drought risk perception and behavioral responses to adapt to drought. Both chapters provide information of research context, research methods, and the results of descriptive and regression analyses. These are followed by several policy implications and future research directions, along with potential limitations with methodology. The sixth chapter qualitatively approaches the research question from an expert’s perspective. The key point here is to fill the gaps between the hypotheses and the findings of the two quantitative analyses. The components of this chapter are similar to those of the previous two chapters.
Synthesis and conclusion. The seventh chapter briefly summarizes the content of the main chapters (from II to VI), including the results of the three empirical analyses, and evaluates the attainment of study objectives. Finally, it presents useful lessons for policymakers and researchers respectively, and ends with a short epilogue.
6


CHAPTER II
BACKGROUND Fundamentals of Climate Change
Despite the efforts of the Intergovernmental Panel on Climate Change and the international community, the outlook of future international climate policy remains bleak. Beginning of Anthropogenic Climate Change
Since the existence of humankind, the interaction between human and environment has ensued in direct or indirect ways. During this interaction process, human activities often caused a disturbance of natural cycles that resulted in human misery (Dankelman, 2010). Severe floods and desertification caused by deforestation of mountains (i.e., the overuse of natural resources) in human history might be good cases in point (Dankelman, 2010). Climate change brought by global warming is somewhat similar to these disasters in that human activities, called Greenhouse Gas (GHG) emissions, are responsible for it.4 However, climate change should not be treated equally as national or local disasters (e.g., earthquakes, wildfires, and hurricanes). This is because it can trigger social, economic, and environmental damages (e.g., extreme weather events, food shortage, biodiversity loss, and sea level rise) not only on a national or local level, but also on a global level (Bellard et al., 2012; Fischer et al., 2005; Pearce et al., 1996).
More than a century ago, before the use of the term, ‘climate change,’ the Swedish chemist Svante Arrhenius investigated how atmospheric carbon dioxide is connected with air temperature to account for the existence of ice ages (Arrhenius, 1896). In 1964, the possibility of catastrophic climate change generated by human activities was cast by social ecologist Murray
4 The earth’s cooling efficiency to space is reduced by increases in the concentration of the greenhouse gas (e.g., carbon dioxide, methane, nitrous oxide, and chlorofluorocarbo). Global warming triggered by this phenomenon leads to unpredictable climate change (Houghton, Callander, & Varney, 1992; Nordhaus, 1991).
7


Bookchin.5 However, this issue raised sporadically by scholars did not receive much attention from the general public and the international community for a long time.
Emergence of a Global Climate Change Organization
In 1987, the word, ‘sustainable development,’ which refers to the “development that meets the needs of the present without compromising the ability of future generations to meet their own needs” (World Commission on Environment and Development, 1987, p. 8), was introduced by the Brundtland Commission’s report (Giddens, 2011). The next year, a significant global effort for combating climate change first appeared. The Intergovernmental Panel on Climate Change (IPCC) was established jointly by the World Meteorological Organization (WMO) and the United Nations Environment Program (UNEP) (IPCC, 2014a). The purposes of this intergovernmental organization are both to assess the causes and impacts of climate change and to come up with more effective measures for addressing this challenge. The findings of the IPCC’s studies have been announced in a 5-6 year cycle, called an IPCC Assessment Report (IPCC, 2014a). Some researchers have identified problems with IPCC studies, such as a lack of systematic empirical research (Vasileiadou, Heimeriks, & Petersen, 2011), the increased mediatization and politicization (Berkhout, 2010), and a monopoly on the provision of international climate policy recommendations (Tol, 2011). Despite such criticisms, the IPCC’s papers have been rated as the most scientifically authoritative reports on climate change and have provided a framework for international negotiations related to climate change (Bolin, 2007;
Choi, 2016; IPCC, 2014b; Zillman, 2007).
5 Bookchin (1964) writes, “The burning of fossil fuels (coal and oil) add 600 million tons of carbon dioxide to the air annually.... Since the Industrial Revolution, the overall atmospheric mass of carbon dioxide lias increased by 25 percent over earlier, more stable levels.... This is growing blanket of carbon dioxide, by intercepting heat radiated from the earth, will lead to more destructive storm patterns and eventually to melting of the polar ice caps, rising sea levels, and inundation of vast land areas” (p. 22).
8


Recently, the fifth assessment report points out that the main cause of global warming is GHG emissions from human activities, and human and natural systems are facing the greatest crisis due to climate change (IPCC, 2014b). According to the findings of the report, the average global temperature has risen by 0.89 degrees Celsius (0.69-1.08 degrees Celsius) for the past 112 years (1901-2012) due to global warming. Melting glaciers and sea level rise have been evident. Arctic glaciers will continue to recede and Antarctic glaciers have also declined in recent years (2002-2011). As a result, the mean sea level in the 20th century rose 1.7 (1.5-1.9) millimeters per year. Since 1993, it is estimated that sea level rise is accelerating by 3.2 (2.8-3.6) millimeters per year. In other words, climate change is getting faster.
Human-caused GHG emissions are affected by several factors such as an economic activity, energy use, population size, technology, and climate policy. The fifth assessment report provides the standard set of future climate change scenarios, called the Representative Concentration Pathways (RCPs) (IPCC, 2014b). Simply, the RCPs refer to “time-dependent projections of atmospheric greenhouse gas (GHG) concentrations” (Moss etal., 2008, p. 5). The four RCPs described in the fifth assessment report consist of one stringent mitigation (RCP2.6), two intermediate (RCP4.5 and 6.0), and one high GHG emission (RCP8.5) scenarios.6 According to IPCC (2014b), if the international community does not make additional efforts to control GHG emissions, the pathway will be located between the RCP6.0 and RCP8.5. Given the RCP8.5, the global average temperature and sea level at the end of this century (2081-2100) are expected to rise 3.7 degrees Celsius and 63 centimeters, respectively. However, if substantial net
6 Each RCP presents a quantitative description of climate change pollutant (e.g., CO2 and CH4) emissions and concentrations as well as radiative forcing in 2100 (IPCC, 2014b). Radiative forcing, measured in watts per square metre (W/m2), refers to “the additional energy taken up by the Earth system due to the enhanced greenhouse effect” (Bjomtes, 2013, p. 2). The numbers of RCPs mean their radiative forcings by 2100 (see Bjomtes (2013), Moss el at. (2008), or Wayne (2013) for more information).
9


negative GHG emissions that follow the RCP2.6 are realized by the international community, the rise of the global average temperature and sea level at the end of this century can be constrained by 1.0 degrees Celsius and 40 centimeters, respectively. Finally, the report warns that if GHG emissions continue (i.e., the efforts of GHG reduction fail), and the global mean temperature exceeds 2 degrees Celsius above pre-industrial temperatures, damages due to climate change, such as the increase of food shortage, the intensification of terrestrial species’ extinction, and the rise of the death toll in extreme weather events, will most likely spread across the planet during and beyond the 21st century (IPCC, 2014b).
The IPCC has highlighted climate change as a universal issue, and its assessment reports have presented the major risks of climate change, which can have an irreversible impact on humankind (Ford et al., 2016; Zillman, 2007). Also, the IPCC has contributed to creating robust consensus that not only can effective global actions significantly reduce the levels of climate change risks, but they also lead to sustainable development (Choi, 2016). It is true that the IPCC has shown science-based leadership in the climate change field. However, the IPCC is not a real actor on the climate change stage, but an assistant. It cannot solely combat climate change without active players, such as national and local governments and the general public.
Global Efforts for Addressing Climate Change
Since the 1990s, the international community has made efforts and attempts to combat climate change. However, the world is in a precarious situation and needs more efforts. In 1992, the efforts of the international community to address climate change began in earnest as the United Nations Conference on Environment and Development (UNCED) adopted the United Nations Framework Convention on Climate Change (UNFCCC) in Rio, Brazil (Aldy & Stavins, 2012). This convention was based on the scientific results of climate change risk coming from
10


the IPCC and aimed at stabilizing the concentration of GHG in the atmosphere. The principles of the UNFCCC were “common but differentiated responsibilities and respective capabilities (CBDRC),” and these served as the core guideline of climate change negotiations (Brunnee & Streck, 2013, p. 589).
With the UNFCCC as a beginning, the efforts of the international community to prevent severe damages triggered by climate change gradually have expanded. Since 1995, parties to the UNFCCC have held the Conference of the Parties (COP) once a year for an international agreement on GHG emission control (Bodansky, 2001). In 1997, the third conference (COP-3) adopted the Kyoto Protocol, which specified global actions for responding to climate change.
The Kyoto Protocol prescribes binding GHG reduction goals for thirty-eight advanced countries (e.g., France, Germany, and Japan) which is to cut GHG emissions by 5.2 percent compared to the 1990 level during the first commitment period (2008-2012) (Meyerson, 1998). However, doubts about the effectiveness of the Kyoto Protocol arise among the targeted nations. This is because although some countries, such as China, South Korea, and India, have increased GHG emissions along with the economic growth, they are not obliged to cut GHG emissions (Choi, 2016; Helm, 2008). The commitment of the Kyoto Protocol weakens further as the U. S., one of the most powerful countries and the world’ second-largest GHG emitter, refuses to ratify the Kyoto Protocol (Freedman & Jaggi, 2011).
To prepare for the post 2012 climate system after the Kyoto Protocol, in 2007, the COP-13, held in Bali, Indonesia, adopted the Bali Roadmap. Based on this roadmap, the COP-15, held in Copenhagen, Denmark, attempted to draw up a new agreement, which forces both advanced and the rest of countries to participate in the post 2012 climate system, but this conference ended without big progress (Parker et al., 2012). In 2010, the efforts of the international community to
11


address climate change came to first fruition at the Cancun conference (COP-16). Countries participating in the conference agreed to put joint efforts into realizing a long-term goal (i.e., limiting the average temperature rise to within 2 degrees Celsius above pre-industrial levels) through adopting a new agreement (Liu, 2011). Namely, the necessity of a paradigm shift to a low-carbon society was recognized by the international community throughout the COP-16 (Hourcade & Shukla, 2015). In 2013, the COP-19, held in Warsaw, Poland, produced an agreement that each country submits a GHG reduction plan for 2020 and beyond under the slogan “Intended Nationally Determined Contributions (INDCs)” (UNFCCC, 2017).
In 2015, the COP-21, held in Paris, France, adopted the Paris Agreement. It stressed the principle of equality and common but differential responsibilities and acknowledged different national circumstances (Choi, 2016). Through this agreement, the GHG emission reduction system, which was centered on some advanced countries at the Kyoto Protocol, was expanded to all countries. This agreement endorsed by one hundred and ninety-five countries initiated a global action plan to prevent drastic climate change through curbing global warming (European Commission, 2017). From a long-term perspective, it aimed to limit the increase in global average temperature to well below 2 degrees Celsius above pre-industrial levels (Rogelj el al., 2016). To this end, each country should establish its GHG reduction plan along with adaptation plan, and its implementation performance will be checked by the United Nations (UN) (Savaresi, 2016). As of 2016, one hundred and eight-nine countries submitted INDC reports related to the GHG reduction plan to the UN, which account for more than ninety-seven percent of global GHG emissions (Kitous et al., 2016).
Despite the international community’s efforts, the future prospect of climate change at the global level does not seem bright. According to the Climate Action Tracker (CAT), global
12


temperatures will likely increase 3.2 degrees Celsius above pre-industrial levels in 2100 even if major countries abide by their INDC reports (CAT, 2017a).7 8 9 Additionally, the U.S.’s intention to withdraw from the Paris Agreement significantly deteriorates that figure compared to 2.8 degrees Celsius estimated in 2016 (CAT, 2017a). Given that many states and cities in the U.S. have vowed to continue upholding the agreement (Plumer, 2017; Regan, 2017), the real impact of the U.S.’s exit from the deal is likely to be smaller than CAT’s outlook. Nevertheless, it might be true that the withdrawal intention of the U.S. from the Paris Agreement has cast a dark cloud over global collaboration to address climate change (Shear, 2017).
Conditions Surrounding Global Collaboration
Complex characteristics of climate change. Regarding climate change, the Earth, more specifically the atmosphere, seems to have the features of public goods (i.e., non-excludable and non-rivalrous).8 9 These goods (e.g., national security and fresh air) are usually accompanied by free rider problems and insufficient provision (Bodansky, 2012). A free rider refers to a person who gains a benefit without paying all or part of its cost (Gauthier, 1986). Also, negative
7 The Climate Action Tracker, a scientific analysis, is provided jointly by three research organizations such as the Climate Analytics (http://climateanalytics.org/), the Ecofys (https://www.ecotys.com/), and the NewClimate Institute (https://newclimate.org/). It evaluates the efforts of thirty-two nations (e.g., the E.U., China, and South Korea), which cover around eighty percent of global emissions, based on their INDC reports and provides expected global warming consequences over the 21st century (CAT, 2017b).
8 Public goods have two characteristics: Non-excludable and non-rivalrous (Bodansky, 2012; Bueno de Mesquita, 2016; Kaul, Grunberg, & Stem, 1999). The former refers to “my having access to it means you have access to it as well,” and the latter refers to “my using the good does not reduce your access to the good” (Bueno de Mesquita, 2016, p. 105). Recently, the extended concept of public goods, called global public goods, has been studied by many researchers (e.g., Barrett, 2007; Kaul et al., 2003; Kaul, Grunberg, & Stem, 1999). This concept is simply defined as goods that have “non-excludable, non-rival benefits that cut across borders, generations and populations” (Kaul, Gmnberg, & Stem, 1999, p. 452).
9 Ostrom et al. (1999) describe climate change as a new challenge of global commons. According to Ostrom, Gardner, and Walker (1994), goods are categorized into four types (i.e., public goods, common-pool resources, toll goods, and private goods), based on the degrees of exclusion—“the difficulty of excluding individuals from benefiting from a good”—and subtractability—“the subtractability of the benefits consumed by one individual from those available to others” (p. 6). While public goods have difficult exclusion and low subtractability, common-pool resources have difficult exclusion and high subtractability (Ostrom, Gardner, & Walker, 1994). Given the quality of the atmosphere which can be depleted, the perspective of Ostrom et al. (1999) is also right.
13


externalities can be triggered by the excessive use of public goods.10 However, climate change caused by GHG emissions is actually more perplexing and insurmountable than it seems. While some countries emit a greater amount of GHG with economic activities into the atmosphere and it triggers other countries’ social, economic, and environmental losses via global warming, the larger GHG emitters do not compensate them for the losses. Moreover, GHG emissions are likely to be over-supplied. If no one pays the costs (i.e., the reduction of GHG emissions), all countries in one ecosystem, the Earth, will bear immense expenses (i.e., damages from climate change). Those negative impacts can even affect future generations. Looking closer, climate change accompanied by negative impacts across the world and generations should be categorized as a “global public bad” (Kaul, Grunberg, & Stem, 1999, p. 456).11 Unfortunately, negotiations between the party triggering global warming and the party influenced by global warming are hindered by significant challenges such as difficulties in measuring damages and transaction costs (Gruber, 2013; Weimer & Vining, 2011). There is no global organization (or nation) which has legitimacy or coercion related to GHG emissions.12 Further, poor countries are much more vulnerable to damages of climate change than rich countries (Nyong, 2009). The more climate change becomes serious, the more citizens of poor countries will be under the triple torture of
10 Externalities occur “whenever the actions of one party make another party worse or better off, yet the first party neither bears the costs nor receives the benefits of doing so” (Gruber, 2013, p. 122). Negative externalities (e.g., air and noise pollutions) are more problematic.
11 Barrett (1999) also regards Ozone depletion and climate change as global public bads. Similar to public goods, public bads are non-excludable and non-rivalrous (Kolstad, 2000). However, while public goods provide positive utility to consumers and are typically under-supplied, public bads bring disutility to them and are over-supplied (Kaul, Grunberg, & Stem, 1999).
12 Legitimacy means that its members are predisposed to accept the actions of the organization, whether they agree with those actions or not (Bickers & Williams, 2001). Coercion refers to the method of persuading its members to do actions by using force such as imposing fines, taking property, and arresting (Bickers & Williams, 2001). Even though there are several global organizations associated with GHG emissions, such as the IPCC and UNEP, all of them solely depend on agreements ratified by countries. Some of the most powerful countries militarily and economically (e.g., the U.S. and China) are the main emitters of GHG, and global agreements are vulnerable to powerful nations’ withdrawal.
14


abnormal weather conditions, the lack of food and quality water, and social disturbances and economic instability.13 Overall, these characteristics, along with the following unsettled leadership and insufficient technological advancement, are retarding global collaboration for responding to climate change.
Unsettled climate change leadership. Many scholars consider leadership as a core determinant of success or failure in coping with transnational issues (Young, 1991).14 As mentioned above, poor countries, in which a majority of the population derives their livelihood from agriculture or fishery, are likely to bear a larger proportion of climate change’s adverse effects (Nyong, 2009). Also, if the international community does not reduce GHG emissions due to national self-interest, damages of climate change are passed on to future generations in the forms of natural disasters, conflicts of generations, and a larger split between the haves and have-nots. Climate change is not just a political or economic problem. It is deeply connected to ethics and requires the leadership of global leading countries.15 Both the U.S. and China are two of the most powerful countries in various fields as well as the second and first-largest emitters of GHG in the world, respectively (Christoff, 2016). In 2001, Paul G. Harris wrote, “Because the U.S. economy is so large, its diplomatic influence so great and its contributions to environmental
13 The economies of poor countries can easily crash because their industries based on agriculture and fishery are deeply influenced by weather conditions (Fischer et al., 2005). Even though poor countries have contributed little to the intensification of GHG emissions in the atmosphere, they will go through more serious torment than rich countries due to weather disaster-prone industries and the lack of capacity and funding (Fischer et al., 2005; Mendelsohn, Dinar, & Williams, 2006; Nyong, 2009).
14 Considering the work of Northouse (2016), leadership consists of four components (i.e., a process, influence, groups, and common goals) and can be defined as “a process whereby an individual influences a group of individuals to achieve a common goal” (p. 6). In light of Northouse’s definition, climate change leadership might be defined as a process whereby a country (or countries) affects a group of countries to reduce global GHG emissions and minimize the impact of climate change.
15 Previous studies state that ethical leadership can positively affect organizations through reducing counterproductive and unethical behaviors (Mayer et al., 2009), enhancing followers’ moral awareness and judgement (Trevino et al., 1999), and fostering dedication, altruism, and better work willingness of followers (Brown, Trevino, & Harrison, 2005).
15


problems so extensive ... the United States must be part of international solutions to environmental change” (Harris, 2001, p. 34). However, in the same year, the Bush administration refused to ratify the Kyoto Protocol (Oberthiir & Roche Kelly, 2008). This event triggered the schism within the international community. About sixteen years later, the Trump administration once again shocks many countries across the world by announcing the U.S. intention to withdraw from the Paris Agreement (Ellis, 2017; Shear, 2017). In the case of China, even though it has been recognized as one of the leaders in the climate change field by many countries (Parker et al., 2012), China has taken a passive stance and seems unwilling to show strong leadership toward climate change (Harris, 2017; Heggelund, 2007; Lewis, 2008).
In 1999, David A. Heenan and Warren Bennis introduced co-leadership to the climate change field. From their perspective, climate change is a real global issue involved with many tasks and stakeholders, unstable circumstances, and severe damages, and one leader cannot be responsible for this huge and complicated issue. Heenan and Bennis (1999) stress that coleadership should not only permeate all institutions associated with climate change (from the global to local levels), but it also shares values and aspiration to curtail the negative impacts of global warming. Based on their work, Nhamo (2009) provides a co-leadership model for climate change governance (i.e., the leadership vehicle). This model consists of seven leadership zones: “mitigation, adaptation, policy and legal frameworks, technology, financing, social dynamics, and research and development” leaders’ zones (Nhamo, 2009, p. 477). While the co-leadership model can be utilized at various levels ranging from the local to global levels, there is no clear guideline for who mitigation leaders should be, who technology leaders should be, etc. It requires leaders to find their zones by themselves, and this process seems very obscure and subjective. In sum, while leadership can play a pivotal role in enhancing global collaboration to
16


combat climate change, it is still immature. Global leading countries still do not show an example of pursuing the ethical norms for the world in the climate change field. Co-leadership does not seem practical due to the absence of an explicit guideline. Karlsson et al. (2011) write, “It is no easy task to provide effective leadership on such a multifaceted issue which involves high stakes for all parties involved” (p. 104).
Insufficient technological advancement. Increasing efforts to prevent GHG emissions generate increasing economic costs. These huge economic costs have made many countries reluctant to actively participate in global efforts on climate change (Osberghaus et al., 2010). To address this cost problem, the world should make ceaseless efforts to develop long-term and sustainable technologies, instead of depending on short-term and end-of-pipe measures which trigger marginal cost surges (Jaeger et al., 1997).16 17 Such a technological breakthrough cannot be realized immediately, and the longer the international community waits in inventing sustainable technologies, the higher the cost barrier will be (Jaeger et al., 1997). In 1997, Jaeger and colleagues advised that the European Union (E.U.) had better “break through the cost barrier” based on “procedural leadership” rather than “posture as an environmental leader by declaring ambitious targets for short-run emissions reduction” (p. 199). This advice may not be only for the E.U., but be for all developed nations. Actually, many countries have made constant efforts to enhance technologies and experiences related to climate change. For instance, in 2007, the E.U. set the target of 20% renewable energy by 2020, and many of its member states, such as 16 17
16 Long-term and sustainable technologies to curb GHG emissions cannot be predicted in detail. But, considering the work of Jaeger et al. (1997), technologies which can improve energy efficiency, increase the production of renewable energy, produce environmentally benign goods, change people’s consumption patterns, or accelerate ecological innovation might be part of them.
17 End-of-pipe solutions refer to the inhibition of pollution emissions by utilizing additional measures without curtailing resource use (Frondel, Horbach, & Runnings, 2007) or the treatment of the impact on the environment triggered by human activities without adjusting the production process itself (Hammar & Lofgren, 2010).
17


Denmark, Germany, and the Netherlands, have been speeding up the development of clean energy technologies (Lewis & Wiser, 2007; Jacobsson & Lauber, 2006; Lund & Mathiesen, 2009).18 Yet despite many nations’ active efforts, fossil fuel-based industries are still prevalent in the world, and high-level technologies and huge costs linked to climate change mitigation and adaptation strategies are a burden to many technically and economically weak countries. Consequently, the development of inexpensive and easy-to-apply technologies related to climate change can strengthen global collaboration and provides the bedrock of transformation to a low carbon society. But, like leadership, technological advancement is still insufficient to make a breakthrough in enhancing global struggle with climate change.
Necessity of more comprehensive efforts. According to Meijerink and Stiller (2013), “Climate change is undoubtedly one of the most urgent policy issues of our times” (p. 240). The evidence is clear that climate change is occurring and affecting all nations (Brecher, 2015). Abnormal weather conditions (e.g., extremely high temperatures, floods, and droughts), glaciers melting, and sea level rise are evident today. More seriously, it is getting worse (IPCC, 2014b). This is because past and current global efforts, especially the reduction of GHG emissions, are not sufficient to address perilous climate change (Osberghaus etal, 2010). IPCC (2014c) warns that if the rise of the global average temperature exceeds 2 degrees Celsius above pre-industrial temperatures, it will fall into a vicious cycle of raising the temperature, and any human efforts to suppress that will become meaningless. No further fence-sitting can be allowed, and genuine accountability and swift countermeasures are needed. Paradoxically, the characteristics of climate change as a global public bad inform the international community of both the necessities
18 Even though many advanced countries (e.g., the U.S., China, Canada, and Japan) turned away from the Kyoto Protocol (Odeku, Maveneka, & Konanani, 2013), the E.U. remained faithful to its roles as a global leader and has played a significant role in saving it (Oberthur & Roche Kelly, 2008).
18


and the difficulties of global collaboration to address climate change. Leadership and technological advancement do not still create favorable conditions for global efforts to cope with climate change. Under this tough circumstance, to address climate change at the global level, both developed and developing countries should make more efforts to enhance global collaboration and to raise their reduction target of GHG emissions. In addition, considering the fragile global collaboration and the above-mentioned unstable conditions that surround it, this dissertation argues that efforts not only at the national and sub-national levels, but also at the non-governmental level (especially individual level) are indispensable.
Government and Individual’s Responses to Climate Change
To cope with climate change, policymakers must try not only to design and implement effective climate policies, but also to activate individuals ’ voluntary behavioral responses. Anthropogenic-Natural Hybrid Disaster
The identification of climate-related disasters by policymakers, whether they are natural origin or anthropogenic, is one of the key points in shaping response policies. Baum, Fleming, and Davidson (1983) compare the characteristics of natural disasters (e.g., floods, hurricanes, and earthquakes) and those of technological catastrophes (e.g., the collapse of dams and the leakage of toxic waste dumps and radiation). They state that while both of them are sudden, powerful, destructive, and hard to predict, there are two significant differences between them. First, natural disasters are uncontrollable and bring short-lived impacts, while technological catastrophes are more controllable (just a “temporary loss of control in the system”) and pose long-lasting threats (Baum, Fleming, & Davidson, 1983, p. 346). Second, recognizing the guilty party of natural disasters is difficult, whereas the guilty party of technological catastrophes can be more easily identified. This study assumes that while man-made disaster is a broader concept
19


than technological catastrophes, their characteristics are almost identical.19 In the cases of natural disasters which are uncontrollable, have short-lived impacts, and responsible agents are difficult to identify, policymakers may aim to minimize the adverse consequences of these disasters. In terms of man-made disasters which are controllable, have long-lasting threats, and have responsible agents that are more easily identified, policymakers may try to reduce the cause and probability of those disasters through controlling the responsible parties.
The complex identity of climate change does not allow policymakers to simply select one of these dichotomous strategies. The scientific community overwhelmingly has agreed that anthropogenic global climate change is occurring (Oreskes, 2004). The IPCC’s report also indicates that GHG emissions from human activities are the main cause of global warming (IPCC, 2014b). When considering only the cause of climate change, it can be categorized as a man-made disaster and seems controllable. However, climate change is accompanied by natural disasters, commonly called extreme weather events such as floods, droughts, and heat waves.
The pattern of extreme weather and natural hazard events are regarded as a key signature of climate change by scientists (Mirza, 2003; Scheraga & Grambsch, 1998). Swim et al. (2009) also regard these events as “the most salient aspects of global climate change” (p. 78). Further, even climate change is controllable, it takes a great deal of time. In short, climate change has the features of both a man-made disaster and a natural disaster. Climate change policies must deal with both the causes and consequences of this anthropogenic-natural hybrid disaster.20
19 Park (2011) views man-made disasters as catastrophic accidents triggered by “human factors such as poor judgment, poor working condition, poor maintenance of equipment and/or the negligence of the operators” (p. 466). Also, Wiegman and Gutteling (1995) which compare natural disasters with man-made risks echo the conclusion of Baum, Fleming, and Davidson (1983).
20 Since the early 20th century, natural disasters triggered by human activities have been investigated by researcher projects (e.g., Bellegarde-Smith, 2011; Simpson, 1986; Ulrich, 1935). These studies mostly focus on earthquakes.
20


Difficulties in Government Responses to Climate Change
Unfortunately, establishing and implementing effective policies to address climate change is no simple matter for policymakers due to the following three problems.
Uncertainty. The great uncertainty surrounds climate change policy.21 Kane and Shogren (2000) characterize climate change policy as complex and contentious. This is because the relative merits of response policies are liable to alteration in accordance with “one’s view of the path of future technology, estimates of net damages, and adaptability of social and economic systems” (Kane & Shogren, 2000, p. 92). Tompkins and Adger (2005) also note that long time frames, unsure climate change impacts, and unpredictable social and economic futures make it challenging to create climate policies. Policymakers have to design policies, allocate resources wisely to serve the policies, and convince the public to support their decisions under uncertainty.
Balancing responses. Keeping a balance between two response options confuse policymakers. Climate change policies consist of mitigation and adaptation (Fiissel, & Klein, 2006; Osberghaus, et al., 2010; Pielke, 1998; Stem, 2006). Simply, mitigation refers to “actions taken to reduce greenhouse gas emissions and hence to tackle the causes of climate change,” and adaptation refers to “actions taken to deal with the consequences of climate change, both before and after impacts are felt” (Tompkins & Adger, 2005. p. 563).Typically, mitigation brings benefits to all climate-sensitive systems at the global levels and its lead time is decades, whereas adaptation provides benefits to selected systems at the local or regional levels and its lead time is from immediate to decades (Fiissel & Klein, 2006). The formulation of mitigation and adaptation policies has been largely independent (Fiissel & Klein, 2006). Regarding the structure of the
21 Regarding uncertainty, Stem (2006) describes it as “a key element of most aspects of climate change” (p. 9). Deser et al. (2012) also view it as a main challenge in planning climate policies.
21


IPCC, Working Group II and Working Group III deals with adaptation and mitigation, respectively (IPCC, 2017). According to Wilbanks etal. (2003), the perception that the reinforcement of adaptation policies can weaken the requirement for mitigation policies causes an adaptation and mitigation dichotomy. They state that this dichotomy is particularly prevalent where the interests toward mitigation are not strong and the costs of adaptation are considered affordable. Tompkins and Adger (2005) also note that governments make decisions about climate change mitigation and adaptation in different policy domains (e.g., energy sector / nonenergy sector). However, many studies (e.g., Biesbroek, Swart, & Van der Knaap, 2009; Bosello, Carraro, & De Cian, 2010; Jones et al., 2007; Kane & Shogren, 2000; Laukkonen et al., 2009) highlight that the apparent separation of mitigation and adaptation (or trade-offs between them) might result in ineffective climate policies. Namely, the optimal mix of the two options leads to more risk reduction with less costs. Kane and Yohe (2000) write, “Each pursued in a vacuum can easily lead to increased costs of policies and programs with no attendant reduction in climate risk” (p. 2). Overall, policymakers must effectively and efficiently react to climate change, but the difficulty lies in seeking a good balance between mitigation and adaptation.
Prioritization. Climate change is likely to be low on the list of national policy priorities. Leiserowitz (2005) reports that while the environment and climate change have gained a high level of both awareness and concern from the American public, it has not been perceived as an imminent or a high-priority danger by them. In a 2000 Gallup poll, the environment came in 16th place on Americans’ list of the most significant problems confronting the nation today. Among thirteen environmental issues, global warming ranked 12th (Dunlap & Saad, 2001). This phenomenon that the public does not regard climate change as a priority concern is not only confined to the U.S. In 2001, the English public viewed many environmental issues, such as
22


hazardous waste disposal and water and air pollution, as more concerning problems than climate change (Department for Environment, Food and Rural Affairs, 2002). In 2009, global warming and climate change were ranked 13th among sixteen issues for the government (e.g., economy, unemployment, and crime) by the English public (Department for Environment, Food and Rural Affairs, 2009). Bord, O’Connor, and Fisher (2000) note that any environmental issues are less pressing concerns for the public than economic and social issues. Low priority of climate change may cause policymakers’ lack of attention to climate policy and finally lead to belated responses.
From a policymaker’s perspective, uncertainty, standing at the crossroad of mitigation and adaptation, and low policy priority associated with climate change act as barriers to set up optimal climate policies. In reality, it might be challenging for policymakers, especially in technically and economically weak countries, to perfectly overcome these obstacles by themselves. Thus, along with efforts to solve those issues, additional efforts to promote public participation in addressing climate change are necessary. The role of individuals in the climate should not be underestimated. Actually, individual and households, along with particular industries or large organizations, are major sources of climate change. For instance, in the U.S., the amount of energy consumed by individual and households at homes and in nonbusiness travel accounts for around thirty-eight percent of total energy (Gardner & Stem, 2008). Difficulties in Individual Responses to Climate Change
Thankfully, individuals’ behavioral responses to address climate change can have positive effects on climate policies. Previous studies (e.g., Kane & Shogren, 2000; Kane & Yohe, 2000) report that personal adaptive responses should be considered during decision-making related to public investment in mitigation. This dissertation further asserts that individuals’ behavioral responses to cope with climate change are influential factors on both mitigation and
23


adaptation policies, especially where the costs of national-scale mitigation and adaptation are considered insurmountable. The more behavioral responses to cope with climate change individuals take, the less of a burden policymakers will feel in shaping climate policies and allocating limited resources to the policies. However, when focusing just on personal benefits and costs brought by behavioral responses, they are less likely to occur automatically. Recall, governments generally have two options in addressing climate change (i.e., mitigation and adaptation) and implement both mitigation and adaptation policies.22 Meanwhile, individuals can have the third option: No response 23 They can overlook climate change and depend on efforts to address climate change brought by global organizations or governments. This third option is associated with the features of individuals’ mitigation and adaptation responses.
Mitigation response. GHG emissions triggering climate change are derived from fossil fuel energy use, agriculture, and deforestation (Stem, 2006). The first source accounts for around two-thirds of GHG emissions (Stem, 2006). Government policies, such as the transformation from fossil fuel energy to renewable energy and regulation on energy-intensive industries, are indispensable to reduce GHG emissions. However, given that energy is typically used by industries and individuals, individuals’ behavioral adjustments also are needed to mitigate climate change. Semenza et al. (2008) view both government actions and cooperation from consumers as necessary things to mitigate global climate change.
22 Given the work of Smit et al. (2000), mitigation policy is linked to “reducing or stabilizing greenhouse gas emissions or levels,” and adaptation policy is related to “adjustments in ecological-social-economic systems in responses to actual or expected climatic stimuli, their effects or impacts” (p. 225).
23 Actually, no response is not the exclusive option of individuals. Governments also can use this option against climate change. However, even though governments can make no effort to reduce GHG emissions (e.g., the U.S. ’s intention to withdraw from the Paris Agreement), no response to mitigation, they cannot fully avoid the responsibility of adaption related to extreme weather events (i.e., key figures of climate change), such as preventing flood, securing water for drought, and strengthening a response system for heat waves. Overall, no response to climate change is not a realistic option for governments.
24


Individuals’ energy saving activities can vary such as recycling, reducing the amount of waste, using less electric facilities, driving less and using public transportation, and purchasing energy efficient cars (or electric vehicles) and facilities (Bockarjova & Steg, 2014; Masud etal., 2016; O’Connor, Bard, & Fisher, 1999; Rainear & Christensen, 2017; Whitmarsh, 2009a). But, voluntarily taking these actions to mitigate climate change is not a simple task. From a shortterm perspective, benefits and costs are intermingled in human decision making. For instance, in the case of recycling, individuals should prepare spaces for recycling and non-recycling materials separately and invest their time to recycle materials. These actions for recycling are accompanied by tangible costs (e.g., spaces and time) and intangible benefits (e.g., intrinsic satisfactions linked to values).24 In contrast, in terms of using less electric facilities, individuals can obtain tangible financial benefits (e.g., saving money on their electricity bills), whereas they might accept intangible costs (e.g., inconvenience)25 Regarding driving less and using public transportation, while individuals save their money (e.g., gas and parking fees) and keep their health, the trade-off is that they may spend more physical energy (e.g., more walking) and time. Namely, it is uncertain whether individuals’ net utility, individual’s benefits minus direct and indirect costs brought by their behavioral response for responding to climate change, is positive or negative. While those who achieve positive net utility are likely to participate in behavioral responses to mitigate climate change, those who achieve negative net utility are not (i.e., selecting no response option). From a long-term perspective, as the proverb says, “One man sows and another man reaps.” Mitigating climate change has a potential to bring benefits to
24 Many studies report that values (or worldviews), such as altruism and environmentalism, are influential factors on pro-environmental behaviors (Barr, 2007; Clark, Kotchen, & Moore, 2003; Corbett, 2005) or coping actions against climate change (Akerlof et al., 2013; Kahan el al., 2012) (see p. 63 for more information about pro-environmental behaviors and coping actions against climate change).
25 Some of them may get intangible benefits, too (intrinsic satisfaction based on values).
25


world through reducing impacts of climate change. However, even though mitigation actions are conducted by current generations, potential benefits are expected to be enjoyed by next generations. Recall, the lead time of mitigation is decades. While those who value benefits for next generations are likely to participate in those actions to mitigate climate change, those who do not value that for next generations are not (i.e., choosing no response option).
Adaptation response. As mentioned earlier, anthropogenic climate change brings severe extreme weather events, posing a great threat to humankind and ecosystems, such as flood and drought. Extreme weather events themselves are not climate change but natural disasters.26 However, anthropogenic global warming makes extreme weather events more violent and volatile. Thus, extreme weather and natural hazard events getting more intense are viewed as parts of climate change (Mirza, 2003; Scheraga & Grambsch, 1998). Among them, flood and drought are the most common events across the world, and adaptation efforts against these disasters have been made at the national, local, and individual levels. While both flood and drought are related to water (quantity) management, they are quite different from each other.
Flood. According to Berz et al. (2001), “No other event of nature occurs as frequently throughout the world and causes so much damage as flood in all its various forms” (p. 458). Flood, usually led by extreme precipitation, causes extensive damage not only to human life and properties, but also to ecosystems.27 Briefly, it is a matter of uncontrollably excessive water in certain areas. From an individual’s perspective, their adaptation responses toward flood, such as
26 The World Meteorological Organization (2016) defines an extreme weather as “the occurrence of a value of a weather variable above (or below) a threshold value near the upper (or lower) ends (‘tails’) of the range of observed values of the variable” (p. 4).
27 Flood refers to “an overflow or inundation that comes from a river or other body of water and causes or threatens damage” and is triggered by “weather phenomena and events that deliver more precipitation to a drainage basin than can be readily absorbed or stored within the basin” (U.S. Geological Survey, 2017). River flood, flash flood, and storm surge are generally considered as the main types of floods (Barredo, 2007; Berz et al., 2001; Jonkman, 2005).
26


setting up sandbags in front of the house, purchasing flood protection devices (e.g., pumps), and moving expensive furnishings to the second floor (Zaalberg etal., 2009; Grothmann &
Reusswig, 2006), can bring direct benefits (e.g., minimization of property losses and loss of life) to themselves. Numerous research projects (e.g., Lindell & Hwang, 2008; Miceli, Sotgiu, & Settanni, 2008; Takao eta/., 2004; Thieken eta/, 2007; Zaleskiewicz, Piskorz, & Borkowska, 2002) have investigated self-protective behaviors by individuals against flood. Many of them identify that individuals get involved with behavioral responses to address flood based on perceived flood probabilities, perceived risk to life and property, and others (Bubeck, Botzen, & Aerts, 2012). Namely, there is a high possibility that individuals voluntarily take action against flood for obtaining personal benefits, and these activities can help policymakers ease the burdens of climate adaptation policy. Grothmann and Reusswig (2006), which highlight the impacts of individuals’ adaptive behaviors toward flood, note that residents’ self-protective activities at flood-prone urban areas can contribute to the prevention of monetary flood damage by eighty percent and the reduction of the need of public risk management.
Drought. Due to the conflicting concepts held by diverse academic fields, drought, regarded as “one of the deadly natural hazards” (Lloyd-Hughes, 2014, p. 607), has no unanimous definition (Dracup, 1991; Dracup, Lee, & Paulson, 1980; Palmer, 1965; Wilhite & Glantz,
198 5).28 Instead, the simple definitions of Palmer (1965) and Sheffield and Wood (2012) are useful for understanding this event. Palmer (1965) defines drought as a “prolonged and abnormal
28 Drought is usually categorized into three (Dai, 2011; Dracup, Lee, & Paulson, 1980) or four (Thurow & Taylor, 1999; Wilhite & Glantz, 1985) types: a) meteorological drought— a “period of months to years with below-normal precipitation” (Dai, 2011, p. 45); b) agricultural drought—a “period with dry soils that results from below-average precipitation, intense but less frequent rain events, or above-normal evaporation, all of which lead to reduced crop production and plant growth” (Dai, 2011, p. 45); c) hydrological drought—a “period during which streamflows are inadequate to supply established uses under a given water management system” (Linsley, Kohler, & Paulhus, 1975, p. 374); and d) socioeconomic drought—“conditions whereby the water demand outstrips the supply, leading to societal, economic, and environmental impacts” (Mehran, Mazdiyasni, & AghaKouchak, 2015, p. 7520).
27


moisture deficiency” (p. 2), and Sheffield and Wood (2012) describe it as “a deficit of water relative to normal conditions” (p. 11). In brief, drought can be viewed as a water shortage due to less precipitation than normal. Drought has several characteristics. From a general approach based on the work of Wilhite (2000), first, the existence of drought itself is unclear due to the absence of a precise and objective definition of drought. Second, it is uncertain when drought starts and ends because of gradual accumulation and long term persistence of its effects.29 Third, a wider range of geographical areas suffers from drought than other natural disasters (e.g., floods and hurricanes), and the quantification of the loss and necessary relief works is challenging.
From a policy approach based on the work of Kim (2016), first, governments are likely to miss the timing to have a sense of crisis because of drought’s creeping impacts. They expect a water shortage to be solved through one or two rain events. Second, when people or industries suffer from the restriction or interruption of the water supply for a long time due to governments’ belated responses, great social and economic disruption can happen. Third, despite the adverse effects of drought on local societies, drought is underestimated among natural disasters because it hardly brings infrastructure collapsing and casualties.30 Fourth, the severity of drought is determined not only by natural factors (e.g., precipitation), but also by social factors (e.g., dam operation, water infrastructure, and water consumption pattern). Therefore, there can be a great deal of conflict regarding who is responsible for drought.
Given these characteristics of drought, it seems more difficult for individuals to detect and address drought than flood. Also, unlike flood, individuals’ behavioral responses against
29 Zamani, Gorgievski-Duijvesteijn, and Zarafshani (2006) categorize drought into “slow-onset disasters” in opposition to “fast-onset disasters” (p. 679). They present slow evolvement, uncertainty and ambiguity, belated perception, and an unclear low point as the features of the slow-onset disasters.
30 However, given the work of Peduzzi el al. (2003), severe drought in highly vulnerable nations (mostly poor countries) can bring numerous casualties.
28


drought (e.g., saving water) generally bring social benefits rather than personal benefits. While behavioral responses to cope with flood influence personal life and properties (e.g., house and furniture), those against drought affect water resources (e.g., dams, reservoirs, and rivers), regarded as common-pool resources and increase not personal resistance but social resistance to drought.31 Further, whether or not an individual takes water saving actions, he or she can have an equal opportunity of water consumption with others during the period of drought. Namely, individuals are more likely to be reluctant to take action to address drought compared to flood.
In brief, individuals’ behavioral responses can contribute to helping out government climate policies and addressing climate change. However, the process of individuals’ behavioral responses based merely on personal benefits and costs triggered by the responses is not likely to be seamless and reliable.
Advantages of Individuals’ Behavioral Responses to Climate Change
Although the active role of governments is more imperative due to unstable global collaboration associated with climate change, policymakers are also facing real issues in establishing and implementing climate policies, such as uncertainties, keeping a balance between mitigation and adaptation, low policy priority, and a lack of resources which is the most serious problem in developing countries. Many research projects have suggested policy frameworks or specific strategies related to climate change mitigation (Bosetti, Carraro, & Tavoni, 2009; Lutsey & Sperling, 2008) and adaptation (Biesbroek etal., 2010; Lemieux etal., 2014; Lim etal., 2004) to help policymakers. All of those studies are valuable to policymakers to manage climate policies. Especially, Hallegatte (2009) introduces several significant and practical considerations
31 Ostrom, Gardner, and Walker (1994) present two key points of common-pool resources: a) “excluding potential appropriated or limiting appropriation rights of existing users is nontrivial;” and b) “the yield of the resource system is subtractable” (p. 4).
29


to policymakers when handling climate policies: a) the production of benefits, whether or not climate change exists (“no regret strategies”); b) flexibility (“reversible strategies”); c) the reduction of vulnerability with no or low costs (“safety margin strategies”); d) the utilization of institutional or financial tools with technical solutions (“soft strategies”); and others (pp. 243-244). Surprisingly, individuals’ voluntary behavioral responses related to climate change mitigation and adaptation largely meet these options.32
Consequently, curbing the adverse impact of climate change requires not only governments’ policies, but also individuals’ daily actions (Semenza etal., 2008; Shove, 2010). Clayton etal. (2015) stress the importance of individual behaviors: Climate change cannot be effectively addressed without the exploration of its impact on human well-being, public perception, and influential factors on mitigation and adaptation behaviors. Semenza et al. (2008) also write, “Opportunity for immediate action lies in increasing voluntary mitigation by individuals engaging in more sustainable, low-carbon lifestyle choices” (p. 479). Given that individuals’ voluntary behavioral efforts for responding to climate change can support governments’ climate policies and also have the above-mentioned advantages, exploring how to promote individuals’ behavioral responses against climate change (i.e., finding a more robust process of individuals’ behavioral responses instead of depending on the relationship between personal benefits and costs from the responses) and informing policymakers of the findings are significant assignments of the academic world. This study introduces the process of individuals’ behavioral responses to address climate change, mediated by risk perception, in the next chapter.
32 Most individuals’ mitigation or adaptation actions bring themselves benefits such as saving money, improving health, and increasing satisfaction associated with the protection of environment (of course, these actions are accompanied by some costs such as inconvenience and time). Those actions also bring societies benefits, such as saving social resources, regardless of the existence of climate change. Furthermore, their activities are flexible and can be linked to institutional and financial tools.
30


CHAPTER III
THEORY
Risk Perception Driving Behavioral Responses
Despite the strong potential of risk perception in activating individuals ’ behavioral responses to climate change, empirical evidence surrounding this relationship is still insufficient. Climate Change Risk
In light of previous research projects (e.g., IPCC, 2001; Team, Pachauri, & Reisinger, 2007; Whitmarsh, 2009b), climate change is a change in the typical or average weather caused by (direct and indirect) human activities or natural origin.33 However, what inspires human responses to climate change is not climate change itself, but the perceived risk triggered by climate change.34 Climate change risk can be described as the probability of exposure to perilous consequences on the health or living conditions of humans and ecosystems brought by anthropogenic climate change (or conditions led by previous climate change), considering existing studies (e.g., Bohm & Pfister, 2001; Jones, 2001; Pachauri etal., 2014; UN, 1992).35 Recall, the IPCC’s Fifth Assessment Report warns that excessive increases in global mean temperature will prompt several risks such as a decrease in food production, an increase in the risk of extinction of land and freshwater species, and land loss due to coastal flooding (IPCC,
33 Weather and climate are distinct terms. While weather refers to the short-term conditions (changes) of the atmosphere in terms of temperature, precipitation, wind, etc., climate refers to average weather conditions over a long period of time (i.e., a statistical phenomenon) (Lombardi & Sinatra, 2012; National Aeronautics and Space Administration (NASA), 2011; Rebetez, 1996; Weber, 2010; Whitmarsh, 2008).
34 In the environment field, a response refers to “any action taken by any region, nation, community or individual to tackle or manage environmental change, in anticipation of that change or after change has occurred” (Tompkins & Adger, 2005, p. 564).
35 The definition of risk varies depending on the research fields, such as health, economics, and environment (Sjoberg, Moen, & Rundmo, 2004; Yates & Stone, 1992). Simply, risk refers to “the possibility of loss” and has three essential elements, such as “losses, the significance of those losses, and uncertainty associated with those losses” (Yates & Stone, 1992, p. 23). Risk perception refers to “intuitive risk judgments” (Slovic, 2000, p. 220).
31


2014b). These risks threatening human life and ecosystem are accompanied by untold economic damage. Estimates of damage range from US$ 140 billion to US$ 1.4 trillion (IPCC, 2014c). Also, climate change risk has been identified in diverse fields such as human health (Haines et al., 2006; McMichael, Woodruff, & Hales, 2006; Patz et al., 2005), food (Lobell eta/., 2008; McMichael eta/., 2007; Rosenzweig & Parry, 1994), ecosystems (Millar, Stephenson, & Stephens, 2007; Walther eta/., 2002; Winder & Schindler, 2004), and water management (Amell, 2004; Christensen eta/., 2004; Vorosmarty eta/., 2000). Much scientific evidence indicates that human-induced climate change could jeopardize the world’s ecosystems and human societies (Klein, Schipper, & Dessai, 2005; Whitmarsh, 2009a).
Given the prevalence of climate change risk, this points to the question, “to what degree is the general public aware of (or familiar with) climate change risk?” According to Brechin and Bhandari (2011), at least sixty-six percent of the respondents in all the twenty-five countries surveyed (e.g., Germany, South Korea, and the U.S.) believe that the problem of global warming is either very serious or somewhat serious 36 Based on the different cross-national survey data, they note that more than sixty percent of the respondents in all the fifteen countries surveyed (e.g., China., India, and Japan) answer that either people are being substantially affected by climate change now or it will happen within twenty-five years 31 Hagen (2015) reports that more than half of the respondents in the eight countries surveyed (e.g., Germany, Spain, and the U.S.), except the Netherlands, are either highly concerned or concerned about the possible impacts of 36 37
36 More than fifty percent of the respondents in seventeen out of the twenty-five countries answer that the problem of global warming is very serious. The question has a four point scale: very serious, somewhat serious, not too serious, and not a problem. These survey data are derived from the Pew Global Attitudes Project study released in 2009 (Brechin & Bhandari, 2011).
37 Also, a considerable major of the respondents in all the fifteen countries indicate that climate change will have significant human and ecological impacts. The results are based on the data of the World Bank survey administered in 2009 (Brechin & Bhandari, 2011). The findings of other comparative studies related to climate change perception, such as Kim (2011) and Leiserowitz (2007), are in line with those of the above-mentioned studies.
32


climate change. Over sixty-five percent of the respondents in all the nine countries believe that if global warming increases substantially over the next twenty years, either serious negative consequences or moderate negative consequences will take place.38 Over the past three decades, the general public has been exposed to scientific evidence and academic findings crystallizing climate change risk through mass-media, public campaigns, social communities, private communication networks (e.g., family and friends), and others. It is likely to be difficult for the general public to exactly figure out what trajectory climate change follows (will follow) and how much climate change negatively affects (will affect) humanity and ecosystems (in the future) in detail. However, this study asserts that the general public’s grasp of the whole situation has increased. Individuals are increasingly aware that climate change is a serious risk, an on-going risk, and a gradually intensifying risk.
Role of Public Risk Perception
Regarding the fundamental question about what goes on in individuals’ minds and what thought process they usually go through when reacting to climate change, Swim et al. (2009) offer a critical clue. They write, “Human perceptions and judgments about climate change are important because they affect levels of concern and, in turn, the motivation to act” (Swim et al., 2009, p. 37). This dissertation argues that understanding risk perception related to climate change is a prerequisite for revealing the process of individuals’ behavioral responses against climate change. O’Connor, Bard, and Fisher (1999) simply conceptualize climate change risk perception as “perceived likelihood of negative consequences to oneself and society from one specific environmental phenomenon: global warming” (p. 462). Based on existing studies (e.g., Bohm &
38 The survey questions have a five point scale, where 1 is not at all concerned (not likely at all to happen) and 5 is highly concerned (serious negative consequences). The data of cross-national survey are emanated from the “Global Survey on Public Attitudes towards Climate Change research project” initiated in 2010 (Hagen, 2015, p. 68).
33


Pfister, 2001; Sjoberg, Moen, & Rundmo, 2004), this study more precisely defines climate change risk perception as a subjective evaluation of the probability of climate change (or conditions produced by previous climate change) happening and the severity of and vulnerability to its adverse consequences. The role of risk perception is presented in the following paragraphs from two perspectives: A policy perspective and an individual’s behavior perspective.
From a policy perspective. As alluded to previously, the establishment and implementation of effective climate policies at the national and local levels is no simple task, and depending solely on scientific information might not be a desirable way to soften this situation. This is due to the difficulties of conveying complex scientific information associated with global environmental change (e.g., climate change and stratospheric ozone depletion) to policymakers and of making usable information for decision-making purposes (Pielke, 1994). Also, scientific information cannot clearly answer policymakers’ question, “So, what?” or “What does global change mean in my district?” (Pielke, 1994, p. 317). In other words, it is very hard for policymakers to distill practical significance of scientific information related to global environmental change, to generate policy alternatives, and to secure public support.
Instead of scientific information, government policies related to environmental risks are subject to public debate and risk perception, which attract policymakers’ interests (Fischhoff el al., 1981; Johnson & Tversky, 1983). Public risk perception has widely been regarded as a significant factor in pushing or restricting political, economic, and social actions to cope with particular risks (Leiserowitz, 2005). Many researchers also highlight that public risk perception can drive both technological and scientific risk assessments and relevant policies (Correia etal., 1998; Perry, Lindell, & Tierney, 2001). Moreover, understanding public risk perceptions is critical in developing effective policies to alleviate both human-induced adverse impacts on the
34


environment and human vulnerability to climate change impacts (Ruddell et al., 2012). In short, public risk perception plays a significant role in producing effective climate policies and inspiring public support.
From an individual’s behavior perspective. Both nature and society scholars are fundamentally interested in the question: How do people perceive and respond to climate change? (Diggs, 1991). In the health field, in which investigations related to risks and behavioral responses are prevalent, estimating the seriousness of a risk (risk perception), comparing the benefits and costs of responses (coping appraisal), and choosing the response maximizing expected outcomes (behavioral response) are common steps in the process of addressing the risk (Van der Pligt, 1996). Likewise, understanding how people perceive environmental risks is critical to predicting their responses to environmental issues (Fischhoff, 1985). In the climate change field, previous research projects (e.g., Battaglini etal., 2009; Bockaijova & Steg, 2014; Keshavarz & Karami, 2016; Niles, Lubell, & Haden, 2013; Yu etal., 2013) have investigated the relationship between climate change risk perception and individuals’ behavioral responses. Many of them verify that individuals’ risk perceptions of climate change lead to their behavioral responses to address climate change.39
Overall, public risk perception can play a critical role in addressing climate change both from the perspective of policy and that of individuals’ behavior. This study mentioned that individuals are becoming more aware of climate change risk in the previous section. Nevertheless, it may not be easy for the general public to clearly perceive climate change risk
39 A few different point of views also exist. For instance, Mainieri et al. (1997) warn that public perception related to the environment may not perfectly correspond with behaviors. Based on survey data in the Athens area, Greece, Papoulis et al. (2015) report that while the majority of the respondents are concerned about climate change and agree with public actions, they remain unlikely to take individual action to cope with climate change. This is because the respondents believe that climate change should be solved by the efforts of state and industry, which they view as the main stakeholders, rather than those of individuals.
35


due to the three major characteristics of global change: a) the long temporal scales, which hinder humans’ direct perception; b) the extensive spatial scales, which have relatively little connection with everyday life; and c) its inherent abstraction (Kearney, 1994). Also, difficulties of transferring scientific information are likely to hinder the general public from perceiving climate change risk (Rebetez, 1996). Therefore, close attention should be paid to the formation of climate change risk perception as well as the process of individuals’ behavioral responses to address climate change, mediated by that risk perception.
Gaps in Climate Change Risk Perception Literature
Since the early 1970s, when the first Earth Day and the initial peak of the ecology movement emerged (Erskine, 1972), many research projects have investigated public perception regarding environmental issues. Erskine (1972) describes the “unprecedented speed and urgency with which ecological issues have burst into American consciousness” as a “miracle of public opinion” (p. 120). In 1986, Gillroy and Shapiro state that the issue of environmental protection has gained persistent concern in the U.S. The work of Dunlap and Scarce (1991) also supports the evidence of growing public concern toward environmental issues and public support for government actions to address those issues.
Regarding individual behavioral responses, Baldassare and Katz (1992) unveil that those who perceive a serious environmental threat to their health and well-being have more willingness to engage in eco-friendly practices (e.g., recycling and less driving) based on survey data in the U.S.’s Orange County, California. Since the 1990s, research projects related to climate change risk perception also have been increasing with many scientific warnings. The characteristics of the existing studies are as follows:
36


First, initial studies, such as Read etal. (1994), Bostrom etal. (1994), and Berk and Schulman (1995), focus on how people think about climate change and how risk perception affects the public’s willingness to pay to prevent climate change (or support tax-related climate policy). Weber (1997) studies whether farmers’ beliefs in global warming influence behavioral responses based on interview data in the U.S. Later, O’Connor, Bard, and Fisher (1999) explore the relationship between climate change risk perception and willingness both to support government actions (e.g., tax and regulation) and to take voluntary actions (e.g., driving less, avoiding the purchase of a gas guzzler, and using less air conditioning and heat). These topics have been consistently explored by many scholars (e.g., Berk & Fovell 1999; Brouwer, Brander, & Van Beukering, 2008; Veronesi etal., 2014; Viscusi & Zeckhauser, 2006; Zahran etal., 2006). Most of them indicate that the majority of people are concerned about climate change, and risk perception has a positive correlation with willingness to pay.
Second, since the 2000s, many studies have paid attention to the relationship between personal factors and climate change risk perception. These personal factors include past experiences (Keller, Siegrist, & Gutscher, 2006; Slovic etal., 2004), physical vulnerability (Brody eta/., 2008), knowledge (Shi eta/., 2016; van der Linden, 2015), political preferences (Carlton & Jacobson, 2013; McCright & Dunlap, 2011), and diverse socio-demographic characteristics (e.g., gender, age, education, and income). However, this research neglects the linkages between personal factors and behavioral responses addressing climate change.
Third, while research projects about the relationship between risk perception and behavioral responses have been growing recently, most are limited to certain fields, such as farming (Battaglini eta/., 2009; Niles, Lubell, & Haden, 2013; Roco eta/., 2015), stock-farming (Joshi eta/., 2013; Opiyo eta/., 2016), fisheries (Musinguzi eta/., 2016), and tourism (Saarinen
37


& Tervo, 2006), which can be directly damaged by climate change. Unfortunately, the findings of these studies which are mostly related to specific adaptation strategies of certain industries are not necessarily applicable to behavioral responses of the general public.
Fourth, even though some studies (e.g., Deng, Wang, & Yousefpour, 2017; Spence etal., 2011; Whitmarsh, 2008) have explored the causal relationship among personal factors, risk perception, and behavioral responses associated with climate change, these empirical research projects typically employ a single personal factor (e.g., experiences of flood, drought, or climate change), focus on specific industry contexts (e.g., agriculture) and adaptive responses, or utilize the data derived from certain cities and flood and drought-prone areas.
Overall, numerous research projects have studied climate change risk perception. However, their findings do not provide enough evidence about the process of personal factors— risk perception—behavioral responses to mitigate climate change which can be generalized in the majority of people at the national level. Additionally, very few studies (e.g., Deng, Wang, & Yousefpour, 2017) have investigated individuals’ risk perceptions and their behavioral responses to cope with drought, whereas a number of studies (e.g., Lindell & Hwang, 2008; Takao etal., 2004; Thieken etal., 2007; Zaleskiewicz, Piskorz, & Borkowska, 2002) have explored risk perception and self-protective behaviors by individuals against flood.
Filling the Gaps in the Literature: Research Question
Research question. According to Vails and Sarda (2009), “Climate change has already begun, and will go on for decades” (p. 48). For alleviating the adverse effects of global change, human behaviors must be changed (Kearney, 1994). This dissertation also asserts that what national-level policymakers really need is how to encourage people voluntarily participate in behavioral responses to mitigate climate change, and what local level policymakers actually need
38


is how to facilitate people’s adaptive behaviors against to climate change, especially drought. Therefore, to fill the gaps in the literature and support policymakers’ policy-making related to climate change, this study investigates the following research question: How do personal factors and risk perception related to climate change influence personal behaviors to address climate change? This research question is answered in three different contexts: a) climate change risk perception and personal behaviors to mitigate climate change among individuals at the national level; b) drought risk perception and personal behaviors to adapt to drought among individuals at the local level; and c) significant issues influencing the relationship among personal factors, risk perception, and behavioral responses related to climate change as perceived by experts.
Three personal factors. Unlike physical scientists who approach risks from empirical climate observations and models, the risk perceptions of the public are being affected both by scientific and technical descriptions of risk and by a broader spectrum of factors such as psychological, social, and personal experiences (Lorenzoni & Pidgeon, 2006; Slovic, 2000). As mentioned above, many personal factors (e.g., experiences and knowledge) have been utilized by previous research projects. This dissertation focuses on three personal factors: Personal experience, knowledge, and policy assessment. It is important to note that this study does not aim to understand all of the factors that can shape individuals’ behavioral responses to address climate change. Instead, it aims to better understand the connections among these factors, risk perception, and behavioral responses and articulate the role of risk perception as a mediator. While this approach may not capture some of the antecedent variables that are theoretically related to individuals’ risk perception and behavioral responses to climate change, it allows this study to zoom into factors that are both recognized across different literatures as proximal to behavioral responses, and which have policy relevance (also see pp. 61-62).
39


Integrated Research Model
Higher risk perception affected by personal factors (experience, knowledge, and policy assessment) related to climate change may intensify individuals ’ behavioral responses. Theoretical Foundation for Individuals’ Behavioral Responses
Previous research projects investigating the causal relationship among personal factors, risk perception, and behavioral responses associated with climate change lack theoretical, conceptual, and methodological coherence. To enhance the theoretical foundation and to derive hypotheses and a research model, this dissertation introduces three works, Kates (1971), Rogers (1983), and Grothmann and Patt (2005). These three studies provide insights that suggest risk perception acts as a causal mechanism linking personal factors and behavioral responses to address the risk.
First, Kates (1971) proposes a general systems model of human adjustment to natural hazards. The purpose of Kates’ model is to identify key components of the human response system, to present the strengths and direction of the linkages among those components, and to simulate the processes derived from the components and linkages. Simply, he posits that hazard experiences influence risk perception, and risk perception leads to human responses. Figure 1 presents Kates’ general systems model (Kates, 1971, p. 447). Kates’ model hints that individuals’ hazard experiences have a positive relationship with their risk perceptions, and these risk perceptions lead to adjustment strategies. However, it does not specifically describe a process of forming risk perception.40
40 While some natural hazard-related studies (e.g., Hanson, Vitek, & Hanson, 1979; Mitchell, 1974; Pulwarty & Riebsame, 1997) have introduced the work of Kates (1971), no empirical research project related to climate change employed Kates’ general systems model.
40


Figure 1. Kates’ (1971) general systems model of human adjustment to natural hazards Second, Rogers (1975; 1983) introduces Protection Motivation Theory (PMT). He presents the cognitive steps of individuals’ coping action against a risk, which include: a) “the threatened event is severe;” b) “that, oh yes, it can happen to you;” c) “but that there is an effective way to avoid the danger” (Rogers, 1983, p. 153).41 Briefly, the purpose of PMT is to reveal how a threat leads to coping actions (i.e., behavior changes) through protection motivation.42 It has widely been applied in diverse fields, such as health (Floyd, Prentice-Dunn, & Rogers, 2000; Milne, Sheeran, & Orbell, 2000), technology (Ifinedo, 2012; Vance, Siponen, & Pahnila, 2012), and environment (Bockarjova & Steg, 2014; Kim, Jeong, & Hwang, 2013; Rainear & Christensen, 2017) based on broader applicability. Rogers’ (1983) PMT consists of three parts: The acquisition of information, perceptual process, and responses. First, people obtain information from environmental and intrapersonal sources. Second, the acquisition of
41 Rogers (1983) writes these sentences associated with a fear appeal based on the sermon of Edwards (1741), “Sinners in the Hands of an Angry God.”
42 Tunner, Day, and Crask (1989) note that PMT explains “how individuals process threats and select responses to cope with the danger brought about by those threats” (p. 267).
41


information initiates the perceptual process. Throughout this process, people compare rewards with the threat severity and their vulnerability (i.e., the expectancy of being exposed to the threat), called threat appraisal. They also compare self and response efficacy with response cost. This process is called coping appraisal.** Lastly, protection motivation is formed by the results of the two appraisals, and it leads to certain responses. Figure 2 lays out the schema of Rogers’ protection motivation theory (Rogers, 1983, p. 168). PMT also depicts a consecutive process among influential factors, risk perception, and coping actions.
Source of Information
Cognitive Mediating Processes
Environmental
Factors Affecting Response Probability
- Verbal Persuasion
- Observational Learning
Maladaptive
Response
Intrapersonal
- Personality Variables
- Prior Experience
Adaptive
Response
Increasing Decreasing
Intrinsic Rewards Extrinsic Rewards - Severity Vulnerability - Threat Appraisal
Fear arousal )
Response Efficacy Self-Efficacy - Response Costs = Coping Appraisal
C oping Modes
Action or Inhibition of Action
Protection
Motivation
Single Act Repeated Acts Multiple Acts Repeated Multiple Acts
Figure 2. Rogers’ (1983) protection motivation theory
Some research projects applied PMT to natural hazards and climate change. In the case of natural hazards, Mulilis and Lippa (1990) utilized PMT to study whether a negative, threat-inducing message influences home owners’ earthquake preparedness in California, the U.S. Martin, Martin, and Kent (2009) investigated the relationship among direct experience with wildfire, risk perception, and the risk mitigation process based on the data derived from three 43
43 These appraisals can be described by the following questions from an individual's perspective:
(1) Does an individual obtain rewards from the threat? (i.e., perceived rewards)
(2) Does an individual believe that the threat is severe? (i.e., perceived severity)
(3) Does an individual feel vulnerable? (i.e., perceived vulnerability)
(4) Does an individual believe that the adaptive response can be an effective way of avoiding or minimizing the threat? (i.e., response efficacy)
(5) Is an individual confident in his or her capacity to accomplish the adaptive response? (i.e., perceived self-efficacy)
(6) Are the costs related to the adaptive response small and affordable? (i.e., perceived response costs) (Cismaru, Lavack, & Markewitch, 2009; Rogers, 1983).
42


homeowner associations in the western U.S. Poussin, Botzen, and Aerts (2014) applied PMT to flood preparedness in three flood-prone regions in France. Overall, the findings of these studies support Rogers’ (1983) PMT. In terms of climate change, Cismaru el al. (2011) investigated a mix of eleven communication campaigns associated with climate change deriving from English-speaking countries (e.g., “ACT ON CO2” in the United Kingdom (U.K.), “Change.ie” in Ireland, “Climate Change Household Action Campaign - Think Change” in Australia, “Fight Global Warming” in the U.S., and “You Control Climate Change” in the E.U.) (pp. 69-76). The study reports that most PMT variables that affect behavioral changes are utilized by all of the campaigns. Rainear and Christensen (2017) explored whether the five PMT constructs (e.g., severity, vulnerability, and response costs) have a relationship with intentions to engage in pro-environmental behaviors (i.e., mitigating global warming). The findings of this study strongly match PMT.44 Similarly, the work of Kim, Jeong, and Hwang (2013), which examines behavioral intentions related to climate change mitigation among American and Korean students, reports that some components of PMT (e.g., perceived severity and response efficacy) are significant predictors of behavioral intentions to cope with climate change.
Lastly, Grothmann and Patt (2005) develop a process model of private proactive adaptation to climate change based on Rogers’ (1983) PMT. Similar to PMT, This model has two major cognitive processes, called climate change risk appraisal and adaptation appraisal, and each process has two (i.e., “perceived probability and severity”) and three (i.e., “perceived adaptation efficacy, self-efficacy, and adaptation costs”) subcomponents, respectively (p. 204). This model has several features compared with PMT. First, considering the characteristics of
44 While the work of Rainear and Christensen (2017) strongly verities PMT, it has some limitations. Their work focuses on the stage of cognitive mediating processes. The authors do not utilize specific influential factors, such as experiences and knowledge, except socio-demographic characteristics. Also, respondents are limited to a certain job (i.e., students enrolled in a large public university) and a certain area (i.e., a northeastern part of the U.S.).
43


climate change which generally do not bring benefits, this model does not reflect intrinsic and extrinsic rewards in the risk appraisal process. Second, reliance on public adaptation, which negatively affect risk appraisal, is newly added to the model.45 Third, an additional cognitive process, called avoidant maladaptation consisting of fatalism, denial, and wishful thinking, is placed between the appraisals and the adaptation intention (i.e., PMT’s protection motivation)46 This model posits that influential factors, such as personal experiences, public adaptation, and social discourse, have a relationship with risk perception, and this perception formed by the two appraisals influence personal adaptive responses. Figure 3 describes Grothmann and Patt’s (2005) model (Grothmann & Patt, 2005, p. 204).
Figure 3. Grothmann and Patt’s (2005) model of private proactive adaptation to climate change
45 Grothmann and Patf s (2005) posit that successful adaptation policies conducted by public agencies will make people less likely to take action to adapt to climate change.
46 Avoidant maladaptation, also called maladaptive coping or nonprotective responses, lias been operationalized by many previous studies (e.g., Abraham el al., 1994; Bubeck, Botzen, & Aerts, 2012; Fruin, Pratt, & Owen, 1992; Grothmann & Reusswig, 2006; McCrae, 1984). Rippetoe and Rogers (1987, p. 598) provide the definitions of the following core modes: a) avoidance—an “attempt to evade actively or deny the threat;” b) wishful thinking—a “strategy that prompts the use of panaceas or unrealistic solutions to a problem;” c) fatalism—the “acceptance of a stressful situation as unchangeable and complacency in the face of danger because nothing can be done anyway;” and d) hopelessness— “absence of belief in possible solutions to a threat without acceptance.”
44


Research Modeling Approach
It is true that Kates’ (1971), Rogers’ (1983), and Grothmann and Patt’s (2005) models have the potential to analyze the causal relationship among personal factors, climate change risk perception, and behavioral responses to address climate change. However, this study argues that an adjusted research model, which is more practical and tailored to the research question and policymakers’ needs, is necessary when considering the following three important points.
First, diverse personal factors that can be linked to the establishment of climate policy should be included in the research model. Kates’ (1971) model solely focuses on an experience factor. Rogers’ (1983) PMT has both experience and knowledge factors. However, this knowledge factor is limited to observational learning (“seeing what happens to others”) (Rogers, 1983, p. 167).47 Grothmann and Patt’s (2005) model includes experience, knowledge, and policy-related factors.48 While this research project focuses on key personal factors identified by the three works (e.g., experience, knowledge, and policy assessment), these personal factors will be specified to provide multifaceted policy implications based on the findings of previous empirical studies related to climate change risk perception and individuals’ behavioral responses.
Second, cognitive processes connecting personal factors and behavioral responses need to be simplified. Some components in Rogers’ (1983) PMT and Grothmann and Patt’s (2005) model, such as response efficacy, self-efficacy, and response costs, play an important role in deciding adaptive behavioral responses. Generally, adaptive behavioral responses in certain
47 Also, Rogers’ (1983) PMT regards “verbal persuasion (especially fear appeals)” as a factor which initiates the cognitive mediating processes (p. 167). Knowledge related to risks may be transmitted during the process of verbal persuasion. But, this study does not view it as an explicit knowledge factor.
48 Actually, social discourse on climate change risk and adaption is not a personal factor, but it might affect personal knowledge related to climate change. Also, Grothmann and Patt (2005) posit that personal knowledge affects behavioral responses via the adaptation appraisal.
45


fields (e.g., agriculture and stock-farming) can bring both significant costs and benefits directly to people who take action, and a significant level of capacity is required to accomplish those responses. For instance, when farmers suffering from climate change alter their crops, which are less vulnerable to climate change than current ones, they might consider these questions: a) how much can these crops endure climate change and minimize the losses caused by climate change? (i.e., response efficacy); b) is it possible for me to grow these crops? (i.e., self-efficacy); and c) can I afford costs triggered by changing the crop and do the additional benefits derived from altered crops outweigh the costs? (i.e., response costs). However, in the case of the general public’s behavioral responses associated with climate change mitigation or adaptation (especially drought), they typically do not pay high costs and do not receive direct and significant benefits from their behavioral responses (small benefits and costs are often mixed together). High capacity is not necessary for completing behavioral responses. Also, many mitigating activities (e.g., less driving, energy saving, and others) and their significant impacts on the reduction of GHG emissions are well known to the general public, and they are not difficult to practice. Namely, those efficacy-related components seem less important to this study’s research model. Besides, data from two surveys utilized in the following national and local level analyses unfortunately do not include the measurement of efficacy-related components. Thus, the research model excludes the adaptation appraisal and just focuses on the climate change risk appraisal (i.e., climate change risk perception).
Third, the ways in which different personal factors influence behavioral responses to address climate change can vary. The works of Kates (1971), Rogers (1983), and Grothmann and Patt (2005), suggest that personal factors, such as experiences and knowledge, influence risk perception through a cognitive process, and risk perception affects behavioral responses.
46


However, O’Connor, Bard, and Fisher (1999) present a different approach. They do not regard risk perception as an independent cause of environmentally friendly or hostile behaviors, but an essential by-product of environmental beliefs. Namely, O’Connor, Bard, and Fisher (1999) posit that behavioral responses to address climate change are a partial function of climate change risk perception. Some studies pay attention to both a direct path (from personal factors to behavioral responses) and an indirect path (from personal factors to behavioral responses via risk perception). For example, Niles, Lubell, and Haden (2013) investigated the relationship among farmers’ factors, risk perception, and responses in the U.S. They identify both the impact of a direct-path (e.g., from climate change experience to government program participation) and an indirect path (e.g., from climate change experience to government program participation via climate change risk perception). Zaalberg et al. (2009) explored the relationship among past flooding experiences, cognitive appraisals, and coping efforts in the Netherlands. They also verify the direct and indirect impacts through using mediated models predicting coping responses. Therefore, there is a possibility that individuals’ behavioral responses to address climate change are functions of both personal factors and climate change risk perception. To clarify these functions, the research model should be designed to explore the dual path impacts, directly between personal factors and behaviors as well as indirectly via risk perception. Hypotheses and Research Model
Hypotheses. Based on the identified theoretical foundation, research modeling approach, and the findings of existing empirical studies, five hypotheses are proposed.
Experiences. Many studies (e.g., Keller, Siegrist, & Gutscher, 2006; Slovic et al., 2004; Weinstein, 1989) note that risk perception is deeply influenced by past experiences. Regarding climate change, the perception and reaction of the public are largely based on their own
47


experiences, “when they feel that they themselves have been subjected to tangible signs of climate change” (Rebetez, 1996, p. 495). Spence et al. (2011) also stress the importance of the “first-hand experience of its potential consequences” in perceiving and responding to climate change (p. 46). These arguments are theoretically supported by the works of Kates (1971), Rogers (1983), and Grothmann and Patt (2005), which consider personal experiences as one of the key influential factors on risk perception and behavioral responses.
Empirical evidence about the relationship between personal experiences and risk perception related to climate change has been provided by previous studies. This evidence is based on the experiences of certain extreme weather events, such as flood (O’Connor et al.,
2005; Siegrist, & Gutscher, 2006; Whitmarsh, 2008) and drought (Dessai & Sims, 2010; Diggs, 1991), or the experiences of global warming (Akerlof et al., 2013; Donner & McDaniels, 2013; Li, Johnson, & Zaval, 2011). For example, Diggs (1991) studied whether drought experiences affect Great Plains farmers’ perceptions of climate change based on a mail survey in the western North Dakota and northeastern Colorado in 1989. He notes that peoples’ risk perceptions can be solidified by drought experiences. O’Connor et al. (2005) reveal that managers of community water systems who have experience of a flood emergency in the last five years show a higher flood risk perception than managers who do not by analyzing survey data in two eastern American states (South Carolina and Pennsylvania). The work of Akerlof et al. (2013) explored whether perceived personal experiences of global warming (e.g., changes in seasons, weather, and lake levels) have a relationship with risk perception in the U.S.’s Alger County, Michigan. They conclude that respondents who have more perceived personal experience of climate change present higher perceptions of local climate change risks. Based on long-term climate and survey data, Donner and McDaniels (2013) investigated the impact of national temperature fluctuations
48


on individuals’ concerns about climate change in the U.S. The authors report that mean temperature anomalies are correlated with greater concerns about anthropogenic climate change.
While these studies identify the positive correlation between personal experiences and climate change risk perception, Whitmarsh (2008) provides mixed results. He explored how personal experiences of flooding and air pollution affect individuals’ knowledge, attitudes, risk perception, and behavioral responses related to climate change based on interview and survey data in the south of England. The study reports that the experience of floods makes little difference between flood victims’ and non-victims’ risk perceptions of climate change, whereas the experience of air pollution affects air pollution victims’ and non-victims’ risk perception of climate change differently.
The number of studies on the relationship among personal experience, risk perception, and behavioral efforts related to climate change has been growing. For example, O’Connor,
Bard, and Fisher (1999) state that perception related to the bad consequences of climate change has a significant positive relationship with both voluntary actions to mitigate climate change and voting intentions on climate policy proposals. Lubell, Zahran, and Vedlitz (2007) hypothesize that perceived climate change risk positively influences global warming activism (i.e., policy support, political participation, and environmental behaviors). Their empirical test based on telephone survey data in the U.S. successfully verify the hypothesis. Based on survey data in China, Yu et al. (2013) investigated whether public concern about climate change and perception of the impact of climate change affect willingness to take individual actions to address climate change. While public concern about climate change has a significant positive relationship with behavioral intentions, perception of the impact of climate change shows mixed results. Spence et al. (2011) investigated the relationship among personal experience of flooding, concern over
49


climate change, and willingness to save energy based on the data of the national survey across the U.K. in 2010. They reveal not only the positive correlation among them, but also the direct and indirect (via risk perception) impacts of flooding experience on energy-saving behavioral intentions through utilizing path analysis.
Meanwhile, based on structural equation modeling, Deng, Wang, and Yousefpour (2017) explored whether drought experiences influence two perceptions (i.e., climate change and saving water) and water saving behaviors in southwestern China. While path 2 (experience-perception of saving water-behaviors) presents a significant correlation among the three variables, path 1 (experience-perception of climate change-behaviors) shows a significant positive relationship only between perception and adaptive behaviors. In the case of Niles, Lubell, and Haden (2013), they studied the relationship among farmers’ past experience of climate change (i.e., water availability), climate change risk perception, and government program participation through employing path analysis. While the past experience significantly affects the two variables, surprisingly, it has a negative relationship with the government program participation.
Personal experience itself does not ensure a positive relationship with risk perception. This is because each experience is based on the diverse conditions of severity and vulnerability and can bring different subjective appraisals about risks. While experiences of climate change with high severity and vulnerability can present a positive relationship with risk perception, climate change experiences with low severity and vulnerability can have no relation to risk perception and even can have a negative relationship. Also, because risk perception is likely to be shaped by other factors (e.g., core values, education, and gender), the impact of these experiences can be mitigated by these other factors. However, previous studies have found that the impacts of climate change are getting more severe, and many nations and local areas are
50


highly vulnerable to those of climate change. The majority of existing empirical studies identify positive relationships between personal experiences and risk perception associated with climate change.
To better understand the structural relationship among personal experiences, risk perception, and behavioral responses to address climate change, this study divides personal experiences related to climate change into two types: Short-term experience and long-term experience 49 This approach reflects the definition of personal climate change experience described in Appendix A. The short-term experience consists of the personal experiences of extreme weather events (i.e., drought and flood). The long-term experience is composed of the personal experiences of four climate change features (i.e., the rising of average temperature, the change in precipitation, the increase of the danger of flood, and the increase of the danger of drought in the last ten years). While some research projects (e.g., Mase, Gramig, & Prokopy, 2017; Niles, Lubell, & Haden, 2013) view individuals’ statements about climate change features (e.g., the global climate is changing and average global temperatures are increasing) as climate change beliefs, this study considers that the individuals’ statements about climate change features based on a long-term period (e.g., at least a decade) can be treated as experiences.50 Based on the identified theoretical approaches (i.e., Grothmann & Patt, 2005; Kates, 1971; Rogers, 1983) and much empirical evidence, this study posits that personal experiences (both the short-term and long-term experiences) related to climate change have a positive correlation
49 The words, short and long terms, are subjective, and there is no clear guideline about them. Considering the IPCC’s analysis period (more than decades) (IPCC, 2014b), this dissertation views a long-term as at least a decade. Fussel and Klein (2006) also note that the lead time of mitigation is decades.
50 Climate change, which is a long temporal, extensive spatial, and far from our daily environment phenomenon, cannot be easily experienced by individuals (Swim et al., 2011). However, recent studies (e.g., Akerlof et al., 2013; Howe et al., 2013) indicate that climate change in the local level has been fairly accurately detected by individuals. Therefore, this study views the long-term experience as perceived experiences. Nevertheless, it might be true that individuals’ perception of climate change experiences is tightly intertwined with beliefs.
51


climate change risk perception. Additionally, given some studies (e.g., Niles, Lubell, & Haden, 2013; Spence et al., 2011; Zaalberg et al., 2009) have identified the direct impact of past experiences on behavioral responses, along with the indirect impact via risk perception, this study also expects the existence of the direct and significant path impact between the short-term experience and behavioral responses to address climate change. The first hypothesis is proposed: Hypothesis 1: The more experience with climate change people have, the higher level of
climate change risk perception they will have (and in the case of the short-term experience, the more actions to address climate change they will take)51 Additionally, to analyze the relationship among personal experiences, drought risk perception, and adaptive behaviors against drought, experience of water supply problems (e.g., limited water supply, low water pressure, etc.) during a drought period is utilized as an experiential factor.
Knowledge. The knowledge theory, which is the most widely held theory of risk perception, implies that “people perceive technologies (and other things) to be dangerous because they know them to be dangerous” (Wildavsky & Dake, 1990, p.42).52 Rogers’ (1983) PMT posits that cognitive mediating processes (i.e., shaping risk perception), which lead to coping actions, can be initiated by observational learning. Grothmann and Patt’s (2005) model considers knowledge as a significant factor influencing climate change risk perception. Kates’ (1971) model also implicitly indicates that knowledge influences human adjustments to cope with natural hazards.
51 Experience with climate change and climate change risk perception include experience with drought and drought risk perception, respectively.
52 Wildavsky and Dake (1990) investigated the relationship between several influential factors (e.g., knowledge, personality, political orientation, and cultural biases) and risk perception. They identity a minimal relationship between self-rated knowledge and risk perception.
52


Knowledge-related factors have been used in a number of empirical research projects. There is no doubt that knowledge is a potential factor increasing climate change risk perception or behavioral responses to address climate change. For instance, Malka, Krosnick, and Langer (2009) studied the relationship between knowledge and concern about global warming based on data from two surveys in the U.S. They conclude that more knowledge leads to greater concern among Democrats, Independents, and people trusting scientists to convey dependable information. In New Zealand, Milfont (2012) explored the relationship among knowledge, risk perception, and perceived efficacy related to global warming and climate change by utilizing panel data (three time points from 2008 to 2009). The author reports that concerns about global warming and climate change are significantly intensified by knowing more about them.53 Lee et al. (2015) examined the significant predictors of the public’s climate change awareness and risk perception. By analyzing the data of 119 countries, which derived from the Gallup World Poll in 2007-2008, they note that climate change risk perceptions are strongly predicted by the understanding of human-caused climate change, especially in Latin America and Europe. In the same vein, based on online survey data of six countries (e.g., Canada, China, the U.K., and the U.S.), Shi et al. (2016) investigated the relationship between three knowledge factors (i.e., physical characteristics, causes, and consequences) associated with climate change and concern about climate change. The research findings unveil that the knowledge factors, especially the causes of climate change, are important drivers of climate change risk perception.
O’Connor, Bard, and Fisher (1999) use knowledge, measured by the level of understanding about the causes of climate change, as an independent variable when assessing
53 Also, Milfont (2012) stresses that higher concerns about global warming and climate change increase perceived efficacy and responsibility to help address them.
53


voluntary actions to alleviate climate change and voting intentions about climate policy proposals. They find that knowledge has a significant positive relationship with both voluntary actions and voting intentions. Halady and Rao (2010) explored the linkage between awareness to climate change and behavioral changes amongst individual managers in India. They report that behavioral changes of individual managers are significantly influenced by awareness to climate change. In the case of Yu etal. (2013), they categorize the causes of climate change as a perception factor instead of a knowledge factor. The results of their study are consistent with previous studies: The less that people perceive climate change as natural consequences, the more that people take action to cope with climate change. Meanwhile, Lubell, Zahran, and Vedlitz (2007) hypothesize global warming knowledge as a positive factor on global warming activism (i.e., policy support, political participation, and environmental behaviors). Unlike the perceived risk factor, which has a positive relationship with the three activities, they find that global warming knowledge significantly and positively affects only one activity (i.e., policy support). Based on survey data from an Alpine region in Switzerland, Jaeger etal. (1993) also reveal that while the level of educational achievement is a significant predictor of climate-relevant behaviors, knowledge about climatic risks is not.
Considering the findings of both theoretical and empirical studies, knowledge, mostly associated with general features of climate change (e.g., causes and consequences), typically shows a positive correlation with climate change risk perception or coping actions against climate change. The catch here is that many empirical studies employing knowledge-related factors (e.g., Kellstedt, Zahran, & Vedlitz, 2008; Malka, Krosnick, & Langer, 2009; Milfont, 2012; Shi et al., 2016; Stevenson et al., 2014) focus not on behavioral responses to address climate change, but on climate change risk perception. While some studies (e.g., Halady & Rao,
54


2010; Lubell, Zahran, & Vedlitz, 2007; O’Connor, Bard, & Fisher, 1999) verify the statistically significant relationship between knowledge and behavioral responses to address climate change, they do not consider that climate change risk perception can be a significant medium for connecting them (i.e., the process of personal knowledge—risk perception—behavioral responses).54 Also, in light of the identified theoretical foundation (i.e., Grothmann & Patt,
2005; Kates, 1971; Rogers, 1983), knowledge associated with climate change, especially general characteristics of climate change, should be regarded as an influential factor mediated by climate change risk perception on behavioral responses to address climate change.
To investigate individuals’ behavioral responses to mitigate and adapt climate change, this study utilizes four general knowledge factors, which are commonly used by previous studies, and two specific knowledge factors related to government policy. In the case of mitigation, three factors (i.e., awareness of climate change, the understanding of the cause of climate change, and the familiarity with a government GHG reduction plan) are used as knowledge-related factors. In terms of adaptation, the remaining three factors (i.e., awareness of water quantity conditions, the exposure to water saving information, and the familiarity with subsidy policy during a drought period) are utilized. The application of policy-related knowledge can give policymakers an opportunity to witness whether or not the knowledge diffusion related to climate policy facilitates individuals’ behavioral responses to address climate change. However, no research project provides clear evidence about the impacts of climate policy-related knowledge on risk perception or behavioral responses related to climate change. This study asserts that the
54 Both Lubell, Zahran, and Vedlitz (2007) and O’Connor, Bard, and Fisher (1999) utilize climate change risk perception and knowledge as influential factors on behavioral responses to address climate change independently. Their findings reveal that the impact of climate change risk perception is much stronger than that of knowledge on behavioral responses to cope with climate change.
55


familiarity with government policies to address climate change will lead people to support these policies through their behavioral responses to cope with climate change.55
In summary, this research project views general knowledge of climate change and climate policy-related knowledge as influential factors related to climate risk perception and individuals’ behavioral responses to address climate change, respectively. As an exception, the impact of the exposure to water saving information, one of the general knowledge related adaptations, is investigated through both direct and indirect paths. It refers to individuals’ participation in a campaign, publicity, or education related to water saving. Usually, these events not only inculcate people with general knowledge related to drought and action guidelines to save water, but also motivate them to take action. In other words, unlike the other three knowledge factors, this factor is likely to influence both climate change risk perception and individuals’ behavioral responses against drought. Hypothesis 2 and 3 are proposed:
Hypothesis 2: The more general knowledge associated with climate change people have, the
higher level of climate change risk perception they will have (and in the case of the impact of the exposure to water saving information, the more actions to address climate change they will take).
Hypothesis 3: The more policy-related knowledge associated with climate change people have, the more actions to address climate change they will take.
Policy assessment. Unlike personal experience and knowledge, policy assessment has received minimal attention from scholars. Adger et al. (2009) write, “Policies and non-climatic drivers.. .currently play perhaps an even more important role [than climatic drivers] in
55 According to O’Connor, Bard, and Fisher (1999), knowledge can directly influence environmental behaviors when providing cues for appropriate behavioral strategies. Knowledge related to government policies generally give the general public useful clues what it takes in order to address certain risks.
56


influencing adaptive behaviors to climate change” (p. 335). However, only a few empirical studies utilize assessment factors related to government policy. For example, in 1999, O’Connor, Bard, and Fisher used government helpfulness (i.e., the scale for government working well) as an independent variable that affects individuals’ voting intentions to climate policy proposals (e.g., tax and regulation). The findings of the study reveal that the government helpfulness and individuals’ voting intentions have a significant positive relationship. The research framework of Yu etal. (2013) adopts a policy assessment factor, “confidence on government’s response to climate change” (p. 467). This factor is measured by the level of trust in the effectiveness of the government’s responses to climate change. Based on survey data in China, they unveil that the policy assessment factor has a significant positive relationship with individual behaviors to curb climate change. Also, Niles, Lubell, and Haden (2013) studied the relationship among farmers’ past experience with environmental policies, climate change risk perception, and their response (i.e., intention about government program participation). While prior research finds a significant negative relationship between past policy experience (positive policy assessment) and risk perception, and a significant positive relationship between risk perception and government program participation, the direct impact of past policy experience on government program participation is not detected.
From an empirical perspective, O’Connor, Bard, and Fisher (1999) utilize a policy assessment factor, but the authors do not investigate the relationship between government helpfulness and voluntary actions to address climate change. While Yu et al. (2013) identify a significant relationship between a policy assessment-related factor and behavioral responses to address climate change, they do not consider that climate change risk perception can act as a bridge between them. Lastly, Niles, Lubell, and Haden (2013) fail to verify the direct
57


relationship between personal policy assessment and behavioral responses, whereas their findings validate the process of personal policy assessment shaping risk perception, which in turn shapes behavioral responses related to climate change. From a theoretical perspective,
Grothmann and Patt (2005) include the factor, ‘reliance on public action,’ in their process model of private proactive adaptation to climate change. Reliance on public actions has a negative correlation with climate change risk appraisal, which leads to individuals’ adaptive responses against climate change. Namely, they expect that individuals’ risk perception related to climate change will decrease if climate policies implemented by public agencies can successfully address climate change risk.
Given the results of empirical research projects and the theoretical approach of Grothmann and Patt (2005), policy assessment associated with climate change can be regarded as a negative influential factor on climate change risk perception. Four policy assessment factors are utilized to explore the relationship with climate change risk perception. Concerning mitigation, drought policy assessment and flood policy assessment are employed. Regarding adaptation, water supply assessment and water infrastructure assessment are utilized. Hypothesis 4 is proposed:
Hypothesis 4: The lower people’s level of satisfaction with climate policy is, the higher level of climate change risk perception they will have.
Risk perception. Based on the identified theoretical approaches (i.e., Kates, 1971; Rogers, 1983; Grothmann & Patt, 2005) and the findings of numerous empirical studies, the fifth and last hypothesis is proposed:
Hypothesis 5: The higher people perceive climate change risk, the more actions to address climate change they will take.
58


Controls: Socio-demographic characteristics. Many existing studies associated with climate change risk perception and behavioral responses to address climate change include sociodemographic characteristics, such as age, gender, and education, in their research model. While some factors significantly affect risk perception and behavioral responses, other factors do not. Also, no factor always influences the dependent variables. For example, Semenza etal. (2008) reveal that education and age have a significant relationship with behavioral changes to address climate change. In the study by O’Connor, Bard, and Fisher (1999), only gender significantly affects willingness to take voluntary actions to reduce GHG emissions, whereas gender, age, and education significantly influence voting intentions to climate policy proposals. Yu et al. (2013) report that none of their demographic factors (i.e., age, gender, and the length of residence) are significantly correlated with behavioral intentions to address climate change.
Table 1. Significant socio-demographic factors of previous studies
Authors Socio-demographic Factors Affected Factors
O’Connor, Bard, & Fisher (1999) Gender(Female, +)*** Behavioral intentions
Gender(Female, -)*, Age(+)*, Education(+)** Policy preferences
Leiserowitz (2006) Gender(Female, +)*, Member of environmental group s(+)** Risk perception
Race(White, -)*, Education(+)**, Political ideology (Conservative, -)**, Member of environmental groups(+)** Policy preferences
Zahran et al. (2006) Gender(Male, -)**, Education(+)* Policy preferences
Semenza et al. (2008) Age(-)***, Education(+)* Behavioral responses
Keshavarz & Karami (2016) Age(+)**, Income(-)** Behavioral responses
Rainear & Christensen (2017) Race(no-White, +)*, Age(+)** Behavioral intentions
* Significant at .05, ** significant at .01, *** significant at .001; all two-tailed tests.
Note 1. Behavioral intentions and responses mean willingness to take individual actions and self-reported actions to mitigate (or adapt to) climate change, respectively.
2. Policy preferences refer to the extent of supporting climate policies (e.g., taxes and regulations) to curb climate change.
59


Table 1 arrays the significant socio-demographic factors of several studies related to climate change. When testing the five hypotheses, this study reflects five socio-demographic factors in the research model. Considering the commonly-used factors of previous research projects, gender, age, education, and income are selected as control variables. Lastly, the type of a residential area is added to explore the difference between urban and rural areas. While some inter-personal factors, such as individual’s personal networks and political affiliations, could be relevant theoretically in explaining individuals’ behavioral responses, these factors are outside the scope of this study.
Research Model. Consequently, a research model is provided in Figure 4, based on climate change risk perception, three types of personal factors (i.e., experience, knowledge, and policy assessment), and five socio-demographic factors (i.e., gender, age, education, income, and residential area). Additionally, key concepts to test the hypotheses are defined and operationalized in Appendix A.
Figure 4. Research model of factors affecting individuals’ behavioral responses to address climate change
60


Potential Issues in the Methodology
Scope of the research. Many empirical studies have explored various influential factors on climate change risk perception, such as values or worldviews (e.g., egalitarianism, individualism, and fatalism) (Akerlof etal., 2013; Leiserowitz, 2006), media coverages (Leiserowitz, 2004; Lowe et al., 2006), past experiences (Spence et al., 2011), and knowledge (Shi et al., 2016). van der Linden (2015) states that four dimensions (i.e., cognitive—knowledge, experiential—affect and personal experience, socio-cultural—culture, values, worldviews, and social construction, and socio-demographic—gender, race, income, age, education, and political ideology) primarily affect climate change risk perception. While this dissertation covers van der Linden’s three dimensions (i.e., cognitive, experiential, and socio-demographic), socio-cultural dimension, such as values (deep core beliefs) or worldviews about climate change, is also likely to be associated with risk perception or behavioral responses including policy preference, especially based on research in the U.S. (e.g., Akerlof etal., 2013; Kahan etal., 2012; Leiserowitz, 2006; Smith & Leiserowitz, 2012).56 Assessing socio-cultural dimension is outside the scope of this study, which is a possible limitation of the research but a potentially important area for additional study. However, given the context of South Korea which is a relatively small-size, largely homogenous, and same cultural country, it may not be a critical problem, as deep core beliefs around climate change (e.g., belief in climate change) do not vary as much in South Korea as in the U.S., and values/worldviews positions could be more homogenous.57
56 Values (deep core beliefs) or worldviews may even get intertwined with personal factors. For example, individuals’ beliefs could shape what experiences they get or what knowledge they seek, and individuals’ beliefs could be informed by their experience and knowledge. Nevertheless, this dissertation differentiates experience and knowledge from beliefs, and in line with much of the literature.
57 Social norming (e.g., influence of neighbors or friends’ actions) is also viewed by previous studies (e.g., Lo, 2013; van der Linden, 2015) as a potential source of influence on behavioral responses to climate change. The role of this factor will be covered in Chapter VI (see pp. 155-156) by using the dataset of the local-level analysis.
61


Additionally, from a policymaker’s perspective, the scope of this dissertation has a practical advantage. Evidence about the impacts of the socio-cultural factors derived from different contexts (e.g., cultures, countries, or societies) might not be practically useful information in shaping climate policies and encouraging individuals’ behavioral responses to address climate change within their own cultures, countries, or societies. Even if policymakers try to utilize the socio-cultural factors for affecting individuals’ risk perceptions, it is challenging for them to control those factors, and, even if possible, it may take a significant amount of time and effort. Instead, evidence about the impacts of three factors (i.e., experience, knowledge, and policy assessment) and risk perception can be relatively easy-to-apply information in any contexts when addressing climate change. Also, these factors are possibly adjusted by governments’ efforts.
Measure of behavioral responses. Some studies, such as O’Connor, Bard, and Fisher (1999), Yu et al. (2013), and Rainear and Christensen (2017), employ behavioral intentions to measure individual’s behavioral responses. According to Ajzen and Fishbein (1980), behavioral intentions have a correlation with actual behaviors.58 Meanwhile, Grothmann and Patt (2005) hold a different view: Even though people have intentions, they often do not translate their intentions into actual behaviors due to a lack of resources (e.g., time, money, and staying power), which are not considered during the process of forming the intentions. Some others (e.g., Deng, Wang, & Yousefpour, 2017; Keshavarz & Karami, 2016; Semenza el al., 2008) utilize self-reported behavioral responses instead of behavioral intentions. Gifford, Kormos, and McIntyre (2011) warn of the discrepancies between self-reported behaviors and actual behaviors brought
58 Ajzen and Fishbein (1980) present two important factors for the correlation between intention and behavior. First, the measure of intention must be consistent with the “behavioral criterion in action, target, context, and time.” (p. 51). Second, behavior can be predicted by the measure of intention “only if intention does not change before the behavior is observed.” (p. 52). In other words, to obtain an accurate prediction, the measurement of intention should be as close as possible to the behavioral observation within proper behavioral criterion.
62


by several bias-factors such as personal wishes toward strong pro-environmental scores and social desirability. Namely, regarding the reliability of behavioral response data, there can be a difference between self-reported and observed behavioral responses. However, in reality, it is almost impossible to observe individuals’ daily behaviors. Even if possible, collecting data of individuals’ behavioral responses at the national and local levels requires painstaking efforts, time, and money. A survey method has many advantages, such as providing more effective means of social description, acquiring many information from a single survey, and saving considerable time and money, compared to observation (Singleton & Straits, 2017). Overall, despite problems triggered by the bias-factors, this study practically utilizes self-reported behavioral response data based on the survey.
Relationship with pro-environmental action literature. Pro-environmental behaviors, which Kollmuss and Agyeman (2002) define as “behavior that consciously seeks to minimize the negative impact of one’s actions on the natural and built world” (p. 240), have been investigated by many previous studies (e.g., Dahlstrand, & Biel, 1997; Karp, 1996; Kollmuss & Agyeman, 2002; Sia, Hungerford, & Tomera, 1986; Stem, 2000; Vining & Ebreo, 1990). Actually, individuals’ behavioral responses to mitigate and adapt to climate change somewhat correspond to pro-environmental behaviors (e.g., using less energy, eating vegetarian food, and saving water). However, while pro-environmental action literature has generally focused on social norms, environmental values, and personal propensity as influential factors, this study zooms into the influence of three personal factors (experience, knowledge, and policy assessment) and risk perception on behavioral responses to mitigate long-term climate change and adapt to severe drought. Despite some overlap, this research is worth exploring, and its findings could provide insights both climate change and pro-environmental action literatures.
63


CHAPTER IV
NATIONAL-LEVEL ANALYSIS
To explore the relationship among personal factors, climate change risk perception, and behavioral responses to mitigate climate change, this research project analyzes survey data of individuals from across South Korea. Given the context of South Korea which has rapidly achieved economic growth over the past decades and has experienced the increase of GHG emissions, the findings of this analysis can be meaningful not only to South Korea, but potentially many countries with similar socio-economic conditions in managing GHG emissions and shaping climate policies.
Research Contexts
Economic Development
After the Korean War (1950-1953), South Korea has been able to achieve rapid economic growth based on the government’s development-oriented growth plans, strong leadership, a highly educated and hardworking workforce, and large business groups (“chaeboF) (Amsden, 1989, p. 9; Yoo & Lee, 1987). Its rapid economic growth, known as “the miracle of the Han River,” was accompanied by the advancement of heavy chemical industries and manufacturing (Yoo & Lee, 1987, p. 95). Even though the South Korean economy faced a financial crisis in December 1997, the Korean government promptly overcame that crisis in 2001 through painstaking efforts, such as downsizing the government, restructuring economic systems, privatizing several state-owned enterprises, and fostering information technology (Kim & Moon, 2002).59 Since the 2000s, South Korea has been regarded as one of the leading
59 The International Monetary Fund declared a liquidity crisis and asserted that South Korea was experiencing a “systematic breakdown in economic relations” (Crotty & Lee, 2009, p. 149). In the process of overcoming this crisis, large numbers of businesses went bankrupt, and South Korea straggled against extensive national level economic restructuring and severe public sector reform (Ito, 2007).
64


countries in the Information Communication Technology (ICT) (Larson & Park, 2014; Lee & Ungson, 2008). For instance, According to the ICT Development Index released by the International Telecommunication Union, it ranked the 1st and 2nd out of the one hundred and seventy-six countries in 2016 and 2017, respectively (International Telecommunication Union, 2017). This social and economic development helped yield an increase of GDP per capita, which rose from US$944 in 1960 to US$25,459 in 2016 (World Bank, 2018).60 Greenhouse Gas Emissions
Rapid economic growth is likely to trigger the increase of GHG emissions. As of 2014, Korea ranked 4th (around 568 million tonnes) among thirty-five Organization for Economic Cooperation and Development (OECD) member states in terms of the total amount of carbon dioxide (CO2) emissions, and ranked 2nd (around 145 percent) among them in growth of emissions, from 1990 to 2014 (OECD, 2017).61 Similarly, regarding the amount of CO2 emission per capita and its growth rate from 1990 to 2014, Korea ranked 6th and 1st among thirty-five OECD member countries, respectively (OECD, 2017). Overall, while South Korea has enjoyed economic growth, it now faces challenges in alleviating GHG emissions. Literatures related to Climate Change Risk Perception in South Korea
According to the 2013 national survey on environmental conservation conducted by the Ministry of Environment (2014), in South Korea, the general public indicates climate change as the second most serious problem among ten environmental issues (e.g., household waste, toxic materials, and water pollution), following industrial waste. Also, regarding the improvement
60 Data are expressed constant 2010 U.S. dollars.
61 Carbon dioxide is regarded as “one of the major greenhouse gases” (Worrell et al., 2001, p. 304). According to OECD (2017), greenhouse gases include seven gases, such as carbon dioxide (CO2), methane (CH4), nitrous oxide (N20), chlorofluorocarbons (CFCs), and hydrofluorocarbons (HFCs).
65


priority of environmental issues, climate change is ranked third among the ten problems, following the natural environment and water.62 While many Koreans perceive climate change as a serious and urgent environmental problem, there are very few studies that deal with climate change (risk) perception in South Korea. Also, these existing studies mostly focus on particular fields, such as education (Seo etal., 2013), agriculture and fisheries (Kim, Lee, & Jeong, 2014; Park, Lee, & Kim, 2014), and tourism (Ahn, Lee, & Hwang, 2016; Kim & Kim, 2013; Lee, Hwang, & Seo, 2015; Min, Kim, & Kim, 2015), or certain sub-national locations (Choi, Park, & Jeon, 2014; Kim, Kim, & Lee, 2014; Koh & Yi, 2016).63 In other words, there is no clear evidence about the causal relationship among personal factors, risk perception, and behavioral responses related to climate change at the national level. Thus, this study will become a stepping stone for future research investigating climate change risk perception and behavioral responses of the general public in South Korea. Also, the findings of this national-level analysis could help policymakers design effective climate policies in both South Korea and countries with similar socio-economic conditions.
Research Methods
Data Collection
Research tool. Data were collected from a web-based survey implemented in South Korea. Climate change risk perception and climate policy assessment formed by a psychological process are typically invisible. Personal experiences, knowledge, and behavioral responses to
62 Since 1982, the Ministry of Environment has conducted surveys associated with the national consciousness on environmental conservation in South Korea, typically once every five years (e.g., 2003, 2008, and 2013). The respondents of the 2013 survey consist of 1,500 citizens and 200 environmental experts.
63 Kim, Kim, and Lee (2014) investigated public awareness of climate change in Chungnam province, located in the west of South Korea. Choi, Park, and Jeon (2014) studied public perception of climate change in Gangnam-gu,
Seoul metropolitan government, to support municipality’s climate policy.
66


mitigate climate change at the national level cannot be easily measured by experiments and field studies. The majority of empirical research projects introduced in the literature review chapters depend on survey (e.g., questionnaires or interviews) data or secondary survey data. Therefore, this study obtains these personal and hard-to-observe data at the national level from the Korea Water Resources Corporation (K-water)’s survey results. K-water conducted the survey for establishing South Korea’s water policy against climate change.
While Web-based surveys can reduce costs, save time, and offer flexibility in the questionnaire design, this method has a major weakness, called coverage error (i.e., “difference between the target population and the sampling frame”) (Singleton & Straits, 2017, p. 271).64 As of 2016, the proportion of total fixed and wireless broadband subscriptions in South Korea was over 100% (National Information Society Agency, 2016; OECD, 2016), and about 90% of the entire population utilize smart phones (Statistics Korea, 2017a). Also, large scale Web-based survey platforms, such as ‘www.opensurvey.co.kr’ (around four hundred thousand panels) and ‘www.marketlink.co.kr’ (around half million panels), operate in South Korea (Open Survey, 2017; Marketlink, 2017). Overall, South Korea has a favorable environment for a Web-based survey with less coverage error.
Survey design and refinement. Good survey instrument design is one of the key ingredients to the success of survey research (Singleton & Straits, 2017).65 A questionnaire was developed by researchers at the research center for water policy and economy in the K-water institute, who have substantial experience conducting climate (or water) policy-related studies.
64 One approach for solving this drawback is to utilize “probability-based, Web-enabled panels designed to be representative of target population” (Singleton & Straits, 2017, p. 280).
65 Singleton and Straits (2017) go onto say that “no matter how carefully the sample is selected, a sample survey is only as good as the design of the questionnaire or interview schedule” (p. 301).
67


Singleton and Straits (2017), in evaluating drafts of survey questions, highlight that “whether respondents clearly understand and are able to answer them” (p. 336) is a prerequisite for minimizing the risk of study failure. To evaluate and improve the design of the questions, expert reviews of the survey questionnaire were carried out. Based on the results of the expert reviews, the questionnaire was refined. Key questions in the questionnaire reflect the following topics: a) climate change risk perception (e.g., the seriousness of the adverse impacts of climate change); b) experiences of extreme weather events (e.g., experiences of property damage by drought and flood); c) experiences of climate change (e.g., the rising of average temperature, the change in precipitation, and the increase of the danger of flood and drought); and d) behavioral responses to mitigate climate changes (e.g., saving energy, reducing the use of disposable items, and utilizing renewable energy). Appendix B presents the key questions of the questionnaire. Most questions are closed-ended, mainly in the form of a four point scale.66
Representativeness. This self-administered survey was carried out by the online survey platform, ‘www.opensurvey.co.kr,’ on August 10, 2016 and had 1,500 respondents who were randomly selected among its panels. The target population was individuals who reside in South Korea, 51.7 million as of 2016 (Statistics Korea, 2017b). These respondents represent a good cross-section of gender, age, and region in South Korea.67 The socio-demographic characteristics of the respondents compared to those of the South Korean population are shown in Table 2.
66 According to Singleton and Straits (2017), while it is difficult to develop closed-ended questions, they require less respondent effort and facility with words than open-ended questions. Also, they have merits in “classifying respondents with respect to some well-understood attitude or behavior” and “enhancing standardization by creating the same frame of reference for all respondents” (pp. 306-307).
67 This survey appears to oversample more educated people compared to the population. However, the proportion of people with higher education (i.e., undergraduate and above) among the entire population is increasing quickly (e.g., from 40% in 2011 to 47% in 2016) (Ministry of Education, 2018).
68


Table 2. 2016 survey respondents’ socio-demographic characteristics
Classification Respondents South Korean Population
Num. Per. Num. Per.
Gender
Male 755 50.3% 25,827,594 50.0%
Female 745 49.7% 25,868,622 50.0%
Age
(0 -19) - (10,047,206)
20-29 257 17.1% 6,758,428 16.2%
30-39 292 19.4% 7,534,140 18.1%
40-49 343 22.9% 8,797,053 21.1%
50-59 325 21.7% 8,424661 20.3%
60- 283 18.9% 10,134,728 24.3%
Education
Under middle school 32 2.1% 13.0
High school 416 27.8% 40.0
Undergraduate 888 59.2% 47 0
Graduate and above 164 10.9%
Income (Average Monthly Income before Tax) -
/ KRWMillion (US$ 909) KRW 2 mi Hi on (US$ 1,818) KRW3million (US$ 2,727) KRW 4 mi Hi on (US$ 3,636) KRW5million (US$ 4,545) KRW 6mi Hi on (US$ 5,455)< I 221 14.7% -
Region
Seoul Metropolitan Government 322 21.5% 9,930,616 19.2%>
Busan Metropolitan City 105 7.0% 3,498,529 6.8%
Daegu Metropolitan City 72 4.8% 2,484,557 4.8%
Incheon Metropolitan City 85 5.7% 2,943,069 5.7%
Daejeon Metropolitan City 44 2.9% 1,514,370 2.9%
Gwangju Metropolitan City 41 2.7% 1,469,214 2.8%
Ulsan Metropolitan City 33 2.2% 1,172,304 2.3%
Gyeonggi Province 382 25.5% 12,716,780 24.6%
Gangwon Province 41 2.7% 1,550,806 3.0%
North Chungcheong Province 43 2.9% 1,591,625 3.1%
South Chungcheong Province 51 3.4% 2,096,727 4.1%
North Jeolla Province 51 3.4% 1,864,791 3.6%
South Jeolla Province 48 3.2% 1,903,914 3.7%
North Gyeongsang Province 69 4.6% 2,700,398 5.2%
South Gyeongsang Province 89 5.9% 3,373,871 6.5%
Jeju Special Self-governing Province 18 1.2% 641,597 1.2%
Sejong Special Self-governing City 6 0.4% 243,048 0.5%
Total 1,500 51,696,216
Note: The data of the South Korean population are provided by Statistics Korea (2017b)—Gender, Age,
and Region—and Ministry of Education (2018)—Education.
69


Variable Description
Mediating and dependent variables. Climate change risk perception and behavioral responses to address climate change are employed as a mediating (both dependent and explanatory) variable and a dependent variable, respectively.
Climate change risk perception. This study utilizes a single item, the seriousness of climate change, as an indicator of climate change risk perception. According to Rogers’ (1983) PMT, threat appraisal is conducted based on the two following questions: a) does an individual believe that the threat is severe? (i.e., perceived severity); and b) does an individual feel vulnerable? (i.e., perceived vulnerability). This study considers that perceived seriousness (i.e., the quality of worrying because of possible risks) of climate change can embrace both perceived severity and vulnerability of climate change. For example, if an individual perceives the higher level of severity and the lower level of vulnerability related to climate change, the level of his or her perceived seriousness of climate change may be moderate, and vice versa. If an individual perceives the higher levels of both severity and vulnerability associated with climate change, the level of his or her perceived seriousness of climate change also may be higher, and vice versa. Respondents were asked to select their level of seriousness on the adverse impacts of climate change. They presented their level of seriousness for the question on a four point scale, where 1 was strongly serious and 4 was not serious. This study reverses the scale of the question. Thus, higher values indicate greater perception of the seriousness of the problem.
Behavioral responses to address climate change. This paper uses the number of behavioral responses to mitigate climate change as a dependent variable. Considering the definition of behavioral responses to climate change in Appendix A, this variable includes not only direct behavioral efforts to reduce GHG emissions (e.g., saving energy and utilizing
70


renewable energy), but also indirect behavioral efforts (e.g., donations to environmental organizations) to mitigate climate change. Based on a single question, respondents were asked to indicate all activities they were practicing to address climate change among the following eight activities: a) saving energy; b) saving water; c) using public transportation instead of a privately-owned car; d) reducing the use of disposable items; e) utilizing renewable energy; f) enjoying a vegetarian diet instead of a meat diet; g) donations to environmental organizations; and h) voting based on environmental-related election promises. Recycling, which has usually been utilized by previous studies (e.g., Masud etal., 2016; Semenza etal., 2008; Whitmarsh, 2009a), is excluded from the activities because it is required by law in South Korea (Ministry of Government Legislation, 2015a). The number of practicing activities are coded on a nine point scale (from 0 to 8).68 Additionally, the practice of each activity is employed as a sub dependent variable to reveal more details of the relationship between risk perception and behavioral responses. The answers of the respondents are divided by each activity, and the practice of each activity is dichotomously coded, where 0 = no and 1 =yes.
Independent variables. Three experience variables, three knowledge variables, and two policy assessment variables are utilized as independent variables.
Short-term experiences. Given that drought and flood are most commonly regarded as significant weather events intensified by climate change, the experiences of drought and flood are employed as the indicators of short-term experiences.69 To clarify the range of drought and
68 While each activity has the potential to mitigate climate change, it is very challenging to compare the impacts of those activities. Therefore, this study posits that they have similar effects on climate change mitigation (i.e., giving the same weight to each response), and it might not be highly problematic.
69 The heat wave is also viewed as one of the significant weather events associated with climate change, (Akompab et ah, 2013; Beniston & Diaz, 2004; Hansen, Sato, & Ruedy, 2012). However, it is not dealt with in this analysis due to a lack of data.
71


flood experiences, this study focuses on experiences of property damage by drought and flood. Respondents were asked to choose their answers among four items, where 1 = no experience, 2 = by drought, 3 = by flood, and 4 = by both drought andflood. Experiences of property damage by drought (hereafter the experience of drought) is measured as a dichotomous variable, where no = 0 and yes = 1 (those who selected 2 or 4). Also, experiences of property damage by flood (hereafter the experience of flood) is coded in the same way, where no = 0 and yes = 1 (those who chose 3 or 4).
Long-term experiences. Climate change has long-term features such as variations in temperature and precipitation. To measure the experiences of those features, respondents were asked to indicate their level of experiences on the rising of average temperature, the change in precipitation, the increase of the danger of flood, and the increase of the danger of drought in the last ten years, respectively. They pointed out their level of experiences for each question on a four point scale, where 1 = definitely, 2 = somewhat, 3 = little, and 4 = not. The scales of the questions are reverse coded, where 0 = not, 1 = little, 2 = somewhat, and 3 = definitely. Those values are summed into a single measure, the index of climate change experiences. Cronbach’s alpha of this index is 0.6982. Also, this study tests whether those four items all fall along one dimension. According to principal component factor analysis (the critical point of Eigenvalue is 1), the number of retained factor is one. Specifically, the Eigenvalue of the first factor is 2.16. This is fifty-four percent of the total possible of 4. By comparison, that of the second factor is just 0.74. All correlation values are greater than 0.3 except correlation between the rising of average temperature and the increase of danger of flood (0.27). The overall value of Kaiser-Meyer-Olkin (KMO) measure is 0.7384, and it is located in the middling range (0.70-0.79)
72


(Kaiser, 1974).70 Lastly, the result of Bartlett’s test of sphericity is significant at the p<001 level. Overall, those four items are adequate for creating the index of climate change experiences. The range of this index is from 0 to 12. In other words, the higher the index is, the more intense people’s experience with long-term climate change is.
Knowledge. Two general knowledge variables (i.e., awareness of climate change and the understanding of the cause of climate change) and one specific knowledge variable (i.e., the familiarity with a government GHG reduction plan) were measured.71 First, awareness of climate change is simply coded as a dichotomous variable where no = 0 and yes = 1. Second, regarding the understanding of the cause of climate change, respondents were asked to present the contribution of human activities to climate change by numeric scores from 0 up to 100. The former means that climate change has nothing to do with human activities (e.g., GHG emissions), and the latter means that climate change is caused mainly by human activities. Given the scientific evidence identifying human’s GHG emissions as the main cause of climate change, those who indicate higher scores have a higher understanding of the cause of climate change than those who indicate lower scores. Lastly, the reduction of GHG emissions is a core strategy of governments to mitigate climate change. Respondents’ level of familiarity with a government GHG reduction plan was measured on a four point scale, where 1 was precisely and 4 was not. This study reverses the scale of the question: Higher values indicate greater familiarity with the government GHG reduction plan.
70 Kaiser’s (1974) evaluation for the levels of KMO measure values is as follows: unacceptable—0.00 to 0.49; miserable—0.50 to 0.59; mediocre—0.60 to 0.69; middling—0.70 to 0.79; meritorious—0.80 to 0.89; and marvelous—0.90 to 1.00.
71 The understanding of the consequences of climate change is also one of the important components in climate change knowledge (see Appendix A). Unfortunately, there is no proper item for measuring that component in the survey data.
73


Policy assessment. As alluded to above, drought and flood are extreme weather events deeply associated with climate change. Therefore, policies to address these events are important parts of governments’ coping strategies for climate change. To measure respondents’ assessment of drought policy and flood policy, respondents were asked to answer their level of evaluation for each policy on a five point scale, where 1 = very poor, 2 = poor, 3 = fair, 4 = pood, and 5 = excellent72 73
Control variables. Five socio-demographic variables are included to isolate the influence of personal factors on climate change risk perception and behavioral responses to mitigate climate change. First, gender is coded as a dichotomous variable, where 0 = male and 1 =female. Second, this study includes age expressed in years. Third, the level of education is measured on a four point scale, where 1 = under middle school, 2 = high school, 3 = undergraduate, and 4 = graduate and above. Fourth, this analysis includes the level of income (i.e., average monthly income before tax last year) as a seven point scale variable, range from 1 for those earning less than KRW1 million (around US$ 909) to 7 for those earning over KRW 6million (around US$ 5,455). Lastly, residential area is coded on a three point scale, where 0 = metropolitan government or cities, 1 = middle or small cities, and 2 = counties13 Data Analysis
Given that this study focuses on the relationship among personal factors, risk perception, and behavioral responses related to climate change, regression analysis, a “statistical technique
72 Considering the conceptual map in Appendix A, these policies should be specifically evaluated on their necessities, validities, efficiencies, etc. However, in reality, it is difficult for the general public to precisely assess the policies on those points. Also, the survey includes a single-item for each policy assessment. This study assumes that the overall evaluations for each policy can represent an indicator of policy assessment and will not be problematic.
73 In South Korea, the populations of counties, cities, and metropolitan cities are generally less than 100,000, more than 100,000 and less than 1,000,000, and more than 1,000,000, respectively (Ministry of Government Legislation, 2017b; Statistics Korea, 2019).
74


for investigating and modeling the relationship between variables” (Montgomery, Peck, & Vining, 2012, p. 1), is a suitable tool to evaluate data. Among diverse regression methods (e.g., linear, logistic, and structural), this study employs Structural Equation Modeling (SEM) (Goldberger, 1972). While most regression methods allow analysis of the relationship between variables separately, “only one layer of linkages between independent and dependent variables at a time” (Gefen, Straub, & Boudreau, 2000 p. 4), SEM, including confirmatory factor analysis, path analysis, and others, enables to analyze the relationship among all the variables in the model simultaneously (Gefen, Straub, & Boudreau, 2000; Kline, 2015; Schumacker & Lomax, 2004) and to produce better performance in systematic and comprehensive analyses rather than other regression models (Iacobucci, Saldanha, & Deng, 2007). By utilizing SEM, not only can this study conduct simultaneous analysis of the dynamic relationship among the variables, but also identify the direct and indirect effects of the personal factors on climate change risk perception and behavioral responses to mitigate climate change.
This analysis is composed of four phases. First, socio-demographic characteristics and descriptive results are introduced. Second, bivariate correlations between dependent and other variables are computed. Third, prior to SEM analysis, model fit is evaluated based on the values of chi-square statistic (x2) and several fit indices, such as the Root Mean Square Error of Approximation (RMSEA), Standardized Root Mean Squared Residual (SRMR), Comparative Fit Index (CFI), and Tucker-Lewis Index (TLI).74 Lastly, SEM regression analysis is operated by
74 According to Hu and Bender (1999), the primary goals in the application of SEM are “the assessment of goodness of fit and the estimation of parameters of the hypothesized model(s)” (p. 2). There are two popular methods to assess model fit: The chi-square statistic (%2) and fit indices. Many studies (e.g., Hooper, Coughlan, & Mullen, 2008; Hu & Bender, 1999) categorize fit indices into two types: absolute fit indices and incremental fit indices, also known as relative fit indices (McDonald & Ho, 2002). The former, which measures “how well a priori model reproduces the sample data” (Hu and Bender, 1999, p. 2), includes the Goodness-of-Fit Index (GFI; Joreskog & Sorbom, 1984), Root Mean Square Error of Approximation (RMSEA; Steiger & Lind, 1980), Standardized Root Mean Squared Residual (SRMR; Bender, 1995), etc. The Chi-Square (%2) value is also included in this type. The latter, which evaluates “the proportionate improvement in fit by comparing a target model with a more restricted,
75


using STATA’s SEM package and an appropriate estimation method. Pathways in the national-level research model are presented in Figure 5. In addition to the main model, eight sub models that utilize each behavioral response coded dichotomously as a dependent variable are analyzed by Generalized Structural Equation Modeling (GSEM), and the results are reported in the alternative model specifications section.* 75
Figure 5. Pathways in the national-level research model
Before finalizing the research model, several pilot models, including pathways among personal factor variables, were tested. These pilot models included causal relationships among the subset of personal factor variables, such as between short-term experience and policy assessment (e.g., more drought experiences could be negatively associated with the level of
nested baseline model” (Hu and Bentler, 1999, p. 2), consists of the Comparative Fit Index (CFI; Bentler, 1990), Nonned Fit Index (NFI; Bentler & Bonett, 1980), Tucker-Lewis Index (TLI: Tucker & Lewis, 1973), etc.
75 Specifically, these sub models are analyzed by GSEM—logit. GSEM regression analysis in STATA does not provide model fits.
76


drought policy assessment) and between general knowledge and policy-related knowledge (e.g., greater general knowledge of climate change could be positively associated with the level of policy-related knowledge). However, while these models may help understand the relationship among the personal factors, their model fits are very poor and not acceptable.76 The final models therefore exclude pathways that examine the inter-relationships among the personal factor variables in unveiling individuals’ behavioral responses related to climate change.
Analysis Results
Socio-demographic Characteristics
The respondents are from all regions (i.e., seven metropolitan cities, eight provinces, and two self-governing areas) in South Korea. They range in age from 20 to 87 years (M = 44.40, SD = 13.62) and 49.7% are females. More than half have a bachelor’s degree and 10.9% have an advanced degree. Approximately 65.5% earn over KRW 3million (US$ 2,727) and a bit more than a quarter earn over KRW 5million (US$ 4,545) a month.
Descriptive Results
Most of the respondents are aware of climate change (M = 0.96, SD = 0.21) and view human activities as the key driver of climate change (M = 75.37, SD = 16.36). When asked to consider the impacts of climate change, nearly 98% of the respondents answered that the negative impacts of climate change (M = 3.55, SD = 0.55) are either strongly serious (57.1%) or serious (40.9%). Regarding personal experiences, 11.6% and 21.9% of the respondents have
76 For example, the pilot model adding the pathways between short-term experiences and policy assessment and between general knowledge and policy-related knowledge presents unacceptable model fits (e.g., yl = 1205.787 (df = 35, p<001), RMSEA—0.150, CFI—0.269, TLI—0.253, and SRMR—0.079). Also, while drought experience and awareness of climate change significantly and positively affect drought policy assessment and familiarity with a government GHG reduction plan respectively, flood experience and understanding of the cause of climate change do not have a significant relationship with flood policy assessment and familiarity with a government GHG reduction plan, respectively.
77


experiences of drought and flood, respectively. Meanwhile, the majority of the respondents answer that they (either definitely or somewhat) experience climate changes in the last ten years. As shown in Figure 6, the rising of average temperatures (M = 2.80, SD = 0.45), the change in precipitation (M = 2.52, SD = 0.62), and the increase of the danger of drought (M = 2.55, SD = 0.61) are experienced by more than 90%, respectively. The increase of the danger of flood (M = 2.08, SD = 0.80) is experienced by 75.7%.
Average temperature is rising Precipitation is changing The danger of drought is increasing The danger of flood is increasing
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
â–  Definitely â–  Somewhat Little â–  Not â–  No answer
Figure 6. Results of four long-term climate change experiences
Concerning the government’s flood and drought policies, while 38.1% and 47.7% of the respondents negatively (either poor or very poor) evaluate the flood and drought policies, respectively, only 8.5% and 6.3% of the respondents positively (either good or excellent) assess them, respectively. Many of the respondents are on neutral ground (fair), 53.4% for the flood policy and 46.0% for the drought policy. The proportion of the respondents who are familiar with the government GHG reduction plan (either moderately or precisely) (44.8%) is slightly less than that of the respondents who are not (either hardly or not) (55.2%) (M = 2.33, SD = 0.84). Lastly, as shown in Figure 7, the respondents are practicing around three behavioral responses to address climate change on average (M = 3.49, SD = 1.48).
78


Voting based on environmental-related election promises
Donations to environmental organizations Enjoying a vegetarian diet instead of a meat diet Utilizing renewable energy Reducing the use of disposable items Using public transportation instead of a privately-owned car
Saving water Saving energy
0 300 600 900 1,200 1,500
Figure 7. Results of individuals’ behavioral responses to address climate change Bivariate Correlation Coefficients among Variables
Table 3 presents the inter-correlations of the key variables in the research model. In the case of the relationship between climate change risk perception and independent variables (Table 3.1), all correlations between them show the hypothesized directions (i.e., three experience and two knowledge factors—positive; two policy assessment factors—negative). However, climate change risk perception does not have a statistically significant correlation with short-term experiences, whereas other relationships are significant at the p<001 level. Among sociodemographic variables, gender and age have a significant positive relationship with climate change risk perception at the p<001 and p< 05 levels, respectively. In terms of the correlations between behavioral responses to mitigate climate change and other variables, including climate change risk perception (Table 3.2), the number of behavioral responses are significantly correlated with climate change risk perception and three personal variables at the p<001 level, and all these correlations show positive directions. Among socio-demographic variables, age and income have a significant positive relationship with the behavioral responses at the p<001 and p< 05 levels, respectively.
79


Table 3. Inter-correlations of key variables in the national-level analysis 3.1. Risk perception variable, independent variables, and control variables
Variables Risk perception Experience Knowledge Policy assessment Socio-demographic
Drought Flood CC Awareness Cause Policy Drought Flood Gender Age Education Income Area
Climate change risk perception 1.000
Drought experience 0.034 1.000
Flood experience 0.009 0.337*** 1.000
Climate change experience 0.431*** 0.068** 0.030 1.000
Awareness of climate change 0.138*** -0.023 -0.026 0 ]47*** 1.000
Cause of climate change 0.255*** -0.029 -0.006 0 214*** 0.074** 1.000
Familiarity with a GHG policy 0.109*** 0.195*** 0.120*** 0.224*** 0.131*** 0.008 1.000
Assessment of drought policy -0.108*** 0.064* 0.011 -0.084** -0.013 -0.125*** 0.044 1.000
Assessment of Hood policy -0.092*** 0.064* -0.019 -0.100*** -0.026 -0.110*** 0.051* 0.688*** 1.000
Gender 0.136*** -0.073** -0.090*** 0.083** -0.024 -0.042 -0.146*** 0.011 -0.058* 1.000
Age 0.064* 0.028 0.106*** 0.103*** -0.008 -0.087*** 0.143*** 0.135*** 0.100*** 0.024 1.000
Education -0.001 -0.023 -0.039 0.013 0.069** 0.115*** 0.087*** -0.057* -0.066* -0.161*** -0.137*** 1.000
Income 0.022 -0.035 -0.003 -0.013 0.086*** 0.023 0.076** -0.048 -0.034 -0.057* 0.030 0 979*** 1.000
Area -0.030 0.062* 0.015 -0.032 -0.014 -0.037 0.022 0.014 0.019 -0.025 -0.006 -0.066* -0.077** 1.000
3.2. Behavioral response variable, risk perception variable, independent variables, and control variables
Variables The number of behavioral responses Climate change risk perception Experience Knowledge
Drought Flood GHG reduction plan
Behavioral responses 1.000
Climate change risk perception 0.126*** 1.000
Drought experience 0.104*** 0.034 1.000
Flood experience 0.088*** 0.009 0.337*** 1.000
Familiarity with a GF1G reduction plan 0.166*** 0.109*** 0.195*** 0.120*** 1.000
Gender 0.021 0.136*** -0.073** -0.090*** -0.146***
Age 0 ]97*** 0.064* 0.028 0.106*** 0 ]43***
Education -0.014 -0.001 -0.023 -0.039 0.087***
Income 0.063* 0.022 -0.035 -0.003 0.076**
Area -0.014 -0.030 0.062* 0.015 0.022
*p<.05, **p<.01, ***p<.001 (two-tailed test)


Model Fit
While the Chi-Square value (x2) has traditionally been measured for evaluating model fit for SEM, it has a critical limitation, sensitivity to sample size (i.e., mostly rejecting the model with large samples) (Hox & Bechger, 2007; Hu & Bentler, 1999). Therefore, diverse fit indices, such as Root Mean Square Error of Approximation (RMSEA), Standardized Root Mean Squared Residual (SRMR), Comparative Fit Index (CFI), and Tucker-Lewis Index (TLI), have been proposed by researchers to supplement the Chi-Square (x2) test (Hooper, Coughlan, & Mullen, 2008; Hox & Bechger, 2007; McDonald & Ho, 2002). Regarding the research model, this study reports the results of %2 (Satorra-Bentler scaled) and several fit indices: %2 = 19.122 (df = 6, p<.01), RMSEA = .038 (90% Cl), SRMR = .010, CFI = .970, and TLI = .867. Given acceptable thresholds of fit indices in the previous studies, the research model fits the data well.77 Regression Analysis
This study hypothesizes that personal factors influence climate change risk perception (HI, H2, and H4), and some personal factors and climate change risk perception lead to behavioral responses to address climate change (HI, H3, and H5). SEM is utilized to test the hypotheses about the relationship among personal factor, risk perception, and behavioral responses related to climate change. Prior to analysis, the selection of an appropriate estimation method is needed. Most SEM-related programs set Maximum likelihood (ML) as a default method due to the precision of estimates (smallest variance) (Ullman & Bentler, 2003).
77 Based on previous studies, Hooper, Coughlan, and Mullen (2008)[1] and Hu and Bentler (1999)[2] provide the acceptable thresholds (conventional cutoff values) of fit indices. Those are as follows:____________________
Fit indices Acceptable thresholds (cutoff values)
Chi-square statistic (y2) [1] An insignificant p value (p > .05)
Root Mean Square Error of Approximation (RMSEA) [1] Values less than .07 / [2] less than .06
Standardized Root Mean Squared Residual (SRMR) [1] [2] Values less than .08
Comparative Fit Index (CFI) [1] [2] Values greater than .95
Tucker-Lewis Index (TLI) [1] [2] Values greater than .95
81


However, an important assumption for ML estimation is the full joint normality of all the variables (i.e., multivariate normality) (Chou & Bentler, 1995; Finney & DiStefano, 2006; Ullman & Bentler, 2003; West, Finch, & Curran, 1995).78 Univariate and multivariate normality tests are conducted for variables in the research model. While all univariate skewness and kurtosis of each variable except one variable (i.e., awareness of climate change) are placed beyond the range of non-normality concern (i.e., |skewness| > 2, |Kurtosis| > 7) suggested by previous studies (Chou & Bentler, 1995; West, Finch, & Curran, 1995), the results of Marcha’s multivariate (skewness and kurtosis) test indicate that multivariate normality assumption is not satisfied. Namely, the normality assumption is violated. Considering several approaches for nonnormality (see Footnote 78), the results of SEM regression analyses with different estimation methods (or options) (i.e., ML, ML with robust standard errors, ML with bootstrapping, ADF, and Generalized SEM) are presented in Appendix C. This appendix includes the results of GSEM, specifically Poisson regression which fits models with count outcome variables (Coxe, West, & Aiken, 2009), by assuming that the dependent variable (i.e., the number of behavioral responses) is a count variable. Overall, directions and significance of coefficients are almost identical. Therefore, this study tries to answer the research question by interpreting the results of the SEM regression analysis using ML estimation with robust standard errors.79 The results of
78 When violating the normality assumption, there are several approaches to dealing with non-normality, such as ML with the Satorra-Bentler scaled chi-squares and robust standard errors (Satorra & Bentler, 1988; 1994), bootstrapping (Efron, 1979; Yung & Bentler, 1996; Finney & DiStefano, 2006), and Asymptotically Distribution-Free (ADF) (Browne, 1984). While ADF makes no assumption of normality, it does not show good performance with sample size less than 2,500 (Hu, Bentler, & Kano, 1992; Ullman & Bentler, 2003). Bootstrapping, which refers to “establishing an empirical sampling distribution associated with a statistic of interest by repeatedly sampling from the original parent sample data,” is viewed as a promising method to correct standard errors (Nevitt & Hancock, 2001, p.355). However, it may also present erratic results with small sample size (e.g., 200 or less) (Nevitt & Hancock, 2001). Generally, ML with the Satorra-Bentler scaled chi-squares and robust standard errors is regarded as a good approach to handling non-normality (Curran, West, & Finch, 1996; Hu, Bentler, & Kano, 1992).
79 There is no evidence of multicollinearity, heteroscedasticity, influential outliers, and other significant statistical problems.
82


the SEM regression analysis and significant pathways in the research model are described in Table 4 and Figure 8, respectively. Also, direct, indirect, and total effects of independent variables on behavioral responses to alleviate climate change are presented in Appendix D. Table 4. Results of SEM regression analysis about the relationship among personal factors, risk perception, and behavioral responses related to climate change
Dependent Variables Independent Variables Coeff. Robust Std. Err.
Climate change risk perception
Drought experience .049 .039
Flood experience .001 .032
Climate change experience .108*** .008
Awareness of climate change .216** .079
Understanding of the cause of climate change .005*** .001
Assessment of drought policy -.065** .021
Assessment of flood policy .021 .022
Gender (female) 132*** .026
Age .002 .001
Education -.008 .022
Income .005 .008
Area -.003 .021
Constant 1.830
The number of behavioral responses
Climate change risk perception .260*** .069
Drought experience .300 .156
Flood experience .118 .099
Familiarity with a government GHG reduction plan .206*** .049
Gender (female) .078 .075
Age 017*** .003
Education -.027 .060
Income .044 .024
Area -.036 .063
Constant 1.1 47
n 1,484
Note: Sixteen respondents who did not answer at least one question out of the four questions related to the experiences of climate change are excluded from the SEM regression analysis.
*p<.05, **p<.01, ***p<.001 (two-tailed test)
83


Policy Assessment
*p<.05, **p<.01, ***p<.001 (two-tailed test)
Figure 8. Significant pathways in the national-level research model
Climate change risk perception. Four independent and one socio-demographic variables show a statistically significant relationship with climate change risk perception. Consequently, the results fully identify Hypothesis 2 and partially verify Hypothesis 1 and 4.
Experience. Long-term climate change experiences have a significant positive relationship with climate change risk perception (b = . 108, p<001). This result indicates that the respondents with more intense climate change experiences are likely to have higher levels of climate change risk perception and corroborates the findings of Akerlof etal. (2013), Donner and McDaniels (2013), and Li, Johnson, and Zaval (2011). Contrary to expectation, two short-term experiences do not significantly influence climate change risk perception.
General knowledge. As hypothesized, both awareness of climate change (b = .216) and the understanding of the cause of climate change (b = .005) have a positive correlation with climate change risk perception at the p<01 and p<001 levels, respectively. The more the
84


respondents are aware of climate change or understand the cause of climate change, the higher levels of climate change risk perception they have. These results are not surprising as previous studies (e.g., Lee etal., 2015; Milfont, 2012; Shi etal., 2016) report that more knowledge about climate change, such as the cause and consequence of that, is correlated with greater concern about climate change.
Policy assessment. While this study predicts that the assessments of both drought and flood policies would negatively influence climate change risk perception, only the assessment of drought policy negatively affects climate change risk perception (b = -.065, p<01). The lower respondents’ levels of satisfaction with drought policy are, the higher levels of climate change risk perception they have. This result provides partial evidence to support the argument of Grothmann and Patt (2005) that individuals’ climate change risk perception is expected to increase if climate policies implemented by governments fail to alleviate climate change risk.
Socio-demographic variables. Only gender (b = .132) has a positive relationship with climate change risk perception at the p<001 level. Among the respondents, women have higher levels of climate change risk perception than men. This finding supports other research projects which find that women tend to be more concerned about environmental problems (Blocker & Eckberg, 1997; Bord & O’Connor, 1997; McCright, 2010; Xiao & McCright, 2015).80
Behavioral responses to mitigate climate change. Five independent and two sociodemographic variables directly or indirectly (i.e., via climate change risk perception) affect individuals’ behavioral responses to mitigate climate change. Based on the results of the SEM regression analysis, Hypothesis 3 and5 are fully substantiated.
80 However, even though numerous research projects have investigated gender difference in environmental concern, this topic is still controversial. McCright (2010) goes on to say that “the upshot is a mix of both robust patterns and inconclusive results” (p. 66).
85


Climate change risk perception. This analysis verifies the significant and positive relationship between climate change risk perception and behavioral responses to mitigate climate change (b = .260, p<001). The respondents with greater climate risk perception are more likely to take behavioral responses to alleviate climate change. This finding is consistent with many existing studies, which show climate change risk perception to be a significant motivator of behavioral responses to cope with climate change (e.g., Deng, Wang, & Yousefpour, 2017; Lubell, Zahran, & Vedlitz, 2007; O’Connor, Bard, & Fisher, 1999; Spence et al., 2011; Yu et al., 2013). Long-term climate change experiences, two knowledge variables (i.e., awareness of climate change and the understanding of the cause of climate change), and the assessment of drought policy significantly and indirectly affect behavioral responses to mitigate climate change via climate change risk perception (see Appendix D).
Short-term experiences. Unexpectedly, there are no significant direct effects of short-term experiences on behavioral responses to mitigate climate change. As a result, both drought and flood experiences fail to influence behavioral responses through direct and indirect paths.81 Overall, these findings are not in line with the empirical evidence or arguments of previous studies that show direct experiences of extreme weather events have a positive impact on risk perception or precautionary (or preparedness) actions (Demski etal., 2017; Deng, Wang, & Yousefpour, 2017; Diggs, 1991; O’Connor etal., 2005; Zaalberg etal., 2009). However, this lack of impact from direct experiences is also shown in some other studies (e.g., Carlton et al., 2016; Martin, Martin, & Kent, 2009). This dissertation considers that the relationship between
81 The pilot model, which included three types of short-term experiences (i.e., only drought, only flood, and both drought and flood), showed similar results. No significant relationship between short-term experiences and climate change risk perception and between short-term experiences and behavioral responses to mitigate climate change was found except the weak relationship between both drought and flood experience and behavioral responses to mitigate climate change (b = .434. p<05).
86


Full Text

PAGE 1

ANALYSIS OF THE RELATIONSHIP AMONG PERSONAL FACTORS, RISK PERCEPTION, AND BEHAVIORAL RESPONSES RELATED TO CLIMATE CHANGE IN SOUTH KOREA by YOUN SUK JUNG B.Ec ., Chungnam National University, 2001 M.Ec ., Chungnam National University, 2014 A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Doctor of Philosophy Public Affairs Program 2019

PAGE 2

ii © 2019 YOUN SUK JUNG ALL RIGHTS RESERVED

PAGE 3

iii This thesis for the Doctor of Philosophy degree by Youn Suk Jung has been approved for the Public Affairs Program by Christopher Weible, Chair Tanya He i kkila , Advisor William Swann Kihoon Lee Date: July 27 , 2019

PAGE 4

i v Jung, Youn Suk (P hD, Public Affairs Program ) Analysis of the Relationship among Personal Factors, Risk Perception, and Behavioral Responses related to Climate Change in South Korea Thesis directed by Professor Tanya Heikkila A BSTRACT Despite global collaboration, c limate change is getting worse. Fortunately, i can be alternatives for mitigating Greenhouse Gas (GHG) emissions and alleviating the negative impacts of climate change . However, empirica l evidence about the process of to address climate change, particularly the behavioral responses of the general public, is still insufficient. According to previous studies, climate change risk perception is deeply associated with behavioral responses against climate change . T o strengthen empirical evidence of how behavioral responses to cope with climate change within the general public are shaped , this dissertation investigates the causal relationships among personal factors, risk perception, and behavioral responses related to climate change. It hypothesi zes that personal factors associated with climate change (i.e., experiences, knowledge, and policy assessment) affect climate change risk perception . Subsequently , higher risk perception is likely to intensify behavioral responses to climate change. In addition, some personal factor variables (e.g., extreme weather event experiences and policy behavioral responses. T o assess these hypotheses , both quantitative analyses at the national and loc al levels and qualitative analyses at the expert level are conducted in the South Korea n context. South Korea is contributing to increasing GHG emissions as a result of economic growth over the past decades.

PAGE 5

v First, f or the national level analysis, the measure of behavioral responses focus es on mitigation behaviors (e.g., saving energy and utilizing renewable energy ) . D ata are collected from a web based survey implemented in South Korea ( 2016, N = 1,500) . T he results of regression analysis, specifically Structural Equation Modeling (SEM), are largely consistent s . Most personal factors, such as long term climate change experiences, awareness of climate change , and the assessment o f drought policy, significantly influence climate change risk perception . Subsequently, higher climate change risk perception, along with higher familiarity with a government GHG reduction plan, enhance s behavioral responses to mitigate climate change. Meanwhile , extreme weather event experiences ( e.g., flood and drought) do not have a n impact on either risk perception or behavioral resp onses. Second, t he local level analysis against drought as the measure of behavioral responses . Based on survey data derived from South Ch ungcheong Province , South Korea (2015, N = 515), the SEM regression analysis presents mixed results. S everal personal fac tors, such as exposure to wa ter saving information and the assessment of water in frastructure policy , are significantly correlated with drought risk perception . However , higher d rought risk perception , unlike higher familiarity with a subsidy policy associated with water saving , fails to intensify behavioral responses to adapt to drought. Also, water problem experiences (e.g., limite d water supply) have no relationship with both risk perce ption and behavioral responses. Third, t he last empirical analysis uti lizes expert knowledge from face to face interviews ( 2018, N = 1 3) and an online survey (2018, N = 98) with climate change experts. Their insights offer support for most hypotheses. hyp otheses and the results of the quantitative analyses, the experts recognize that t he significant

PAGE 6

vi relationship between experiences and risk perception requires s everal conditions, such as experiences with direct and significant damage, recent experiences, and instances with no possible solution for minimizing damage . Also, m any of t hem are suspicious about the relationship between drought risk perception and behavioral responses against drought. This is because prioritized residential water suppl y policy and well established res ponse infrastructure generally ensure that individuals are not exposed to direct and sev ere d amage triggered by drought, which make s individuals reluctant to take action. Given the context of South Korea , such as a high level of climate change adaptation capacity, the recent downward trend of damage brought by flood s , and residential water during drought , the identified gaps seem understandable. Overall , while risk perception play s an important role in enhancing behavioral responses to mitigate climate change, the precise role of risk perception in climate change adaptation may depend on the context of each country (e.g., the capacit y of climate change adaptation and recent trends in extreme weather events). D esp ite some limitations , this study successfully offer s empirical evidence of the theorized relationships among personal factors, risk perception, and behavioral responses associated with climate change, which can be important to understand when addressing cl imate change. T he findings of this dissertation and mixed methods research not only contribute to advancing knowledge and the research methodology of the process of be havioral responses to cope with climate change, but they also inform policymakers of critical clues as to how the general public volu ntarily takes action to address climate ch ange . The form and content of this abstract are approved. I recommend its publication. Approved: Tanya Heikkila

PAGE 7

vii TABLE OF CONTENTS CHAPTER I . INTRODUCTION ................................ ................................ ................................ ................ 1 II . BACKGROUND ................................ ................................ ................................ ................ 7 Fundamentals of Climate Change ................................ ................................ .................... 7 Beginning of Anthropogenic Climate Change ................................ .......................... 7 Emergence of a Global Climate Change Organization ................................ ............. 8 Global Efforts for Addressing Climate Change ................................ ...................... 10 Conditions Surrounding Global Collaboration ................................ ....................... 13 Government and Ind es to Climate Change ................................ ........ 19 Anth ................................ ................................ 19 Difficulties in Government Response s to Climate Change ................................ .... 21 Difficulties in Individual Response s to Climate Change ................................ ........ 23 Advantages of Indivi Behavioral Responses to Climate Change .................. 29 III. THEORY ................................ ................................ ................................ ......................... 31 Risk Perception Driving Behavioral Responses ................................ ............................. 31 Climate Change Risk ................................ ................................ .............................. 31 Role of Public Risk Perception ................................ ................................ ............... 33 Gaps in Climate Change Risk Perception Literature ................................ .............. 36 Filling the Gaps in the Literature: Research Question ................................ ............ 38 Integrated Research Model ................................ ................................ ............................. 4 0 Theor e ............................. 4 0 Research Modeling Approach ................................ ................................ ................. 4 5

PAGE 8

viii Hypotheses and Research Model ................................ ................................ ............ 4 7 Potential Issues in the Methodology ................................ ................................ ....... 61 IV. NATIONAL LEVEL ANALYSIS ................................ ................................ .................. 6 4 Research Context s ................................ ................................ ................................ .......... 6 4 Economic Development ................................ ................................ .......................... 6 4 Greenh ouse Gas Emissions ................................ ................................ ..................... 6 5 Literatures related to Climate Change Risk Perception in South Korea ................. 6 5 Research Methods ................................ ................................ ................................ .......... 6 6 Data Collection ................................ ................................ ................................ ....... 6 6 Variable Description ................................ ................................ ............................... 70 Data Analysis ................................ ................................ ................................ .......... 7 4 Ana l ysis Results ................................ ................................ ................................ ............. 7 7 Socio demographic Characteristics ................................ ................................ ........ 7 7 Descriptive Results ................................ ................................ ................................ . 7 7 Bivariate Correlation Coefficients among Variables ................................ .............. 7 9 Model Fit ................................ ................................ ................................ ................. 8 1 Regression Analysis ................................ ................................ ................................ 8 1 Alternative Model Specifications ................................ ................................ ........... 8 7 Policy Implication s ................................ ................................ ................................ ......... 9 2 Limitations and Future Research Directi o ns ................................ ................................ .. 9 5 Conclusion ................................ ................................ ................................ ...................... 9 7 V. LOCAL LEVEL ANALYSIS ................................ ................................ .......................... 9 9 Research Context s ................................ ................................ ................................ .......... 9 9

PAGE 9

ix Water Management and Climate Change ................................ ............................... 9 9 Water Management in South Korea ................................ ................................ ...... 100 Water Condition in South Chungcheong Province ................................ ............... 10 1 Research Methods ................................ ................................ ................................ ........ 10 3 Data Collection ................................ ................................ ................................ ..... 10 3 Variable Description ................................ ................................ ............................. 10 5 Data Analysis ................................ ................................ ................................ ........ 1 10 Ana l ysis Results ................................ ................................ ................................ ........... 1 11 Socio demographic Characteristics ................................ ................................ ...... 1 11 Descriptive Results ................................ ................................ ............................... 1 1 2 Bivariate Correlation Coefficients among Variables ................................ ............ 11 3 Model Fit ................................ ................................ ................................ ............... 11 5 Regression Analysis ................................ ................................ .............................. 11 5 Alternative Model Specifications ................................ ................................ ......... 1 2 2 Policy Implication s ................................ ................................ ................................ ....... 12 3 Limitations and Future Research Directi o ns ................................ ................................ 12 6 Conclusion ................................ ................................ ................................ .................... 12 7 VI. EXPERT LEVEL ANALYSIS ................................ ................................ ..................... 12 9 Research Context s ................................ ................................ ................................ ........ 12 9 Necessity of Mixed Methods Research ................................ ................................ . 12 9 Advantages of Utilizing Experts ................................ ................................ ........... 1 31 Core Components of Experts ................................ ................................ ................ 1 3 2 Expert Interviews ................................ ................................ ................................ ......... 13 3

PAGE 10

x Data Collection ................................ ................................ ................................ ..... 13 3 Data Analysis ................................ ................................ ................................ ........ 13 5 Analysis Results ................................ ................................ ................................ .... 13 7 Expert Survey ................................ ................................ ................................ ............... 15 7 Data Collection ................................ ................................ ................................ ..... 15 7 Socio demographic Characteristics ................................ ................................ ...... 15 7 Analysis Results ................................ ................................ ................................ .... 15 8 Policy Implications ................................ ................................ ................................ ....... 1 6 2 Research Implications ................................ ................................ ................................ .. 1 6 3 Conclusion ................................ ................................ ................................ .................... 16 4 VII. CONCLUSION ................................ ................................ ................................ ............ 16 7 Content of Chapter II VI ................................ ................................ ........................... 16 8 Attainment of Study Objectives ................................ ................................ ................... 1 7 2 Less ons for Policymakers ................................ ................................ ............................. 17 4 Lessons for Researchers ................................ ................................ ............................... 17 5 Epi logue ................................ ................................ ................................ ....................... 17 6 REFERENCE S ................................ ................................ ................................ .................... 17 7 APPENDIX ................................ ................................ ................................ ......................... 2 12 A. Conceptual map for v ariables ................................ ................................ ................. 2 12 B. Key questions of the 2016 s urvey in South Korea ................................ ................. 2 14 C. R esults of SEM regression analyses with different estimation methods (or options) in the national level analysis ................................ ................................ .................. 2 16

PAGE 11

xi D. Direct, ind irect, and total effects of independent and control variables on behavioral responses to mitigate climate change ................................ ................................ ..... 2 17 E. R elationship between personal factors and socio demographic characteristics in the national and local level analyses ................................ ................................ ............ 2 18 F. Key questions of the 2015 s urvey in n orth west area of South Ch ungcheong Province ................................ ................................ ................................ .................. 2 20 G. R esults of SEM regression analyses with different estimation methods (or options) in the local level analysis ................................ ................................ ....................... 2 22 H. Direct, ind irect, and total effects of independent and control variables on behavioral responses to adapt to drought ................................ ................................ ................. 2 24 I. Main questions of the c limate c hange expert interview ................................ ......... 2 25 J. Part of the c odebook for interview transcription ................................ .................... 2 26

PAGE 12

xii LIST OF TABLES TABLE 1 . Significant s ocio demographic factors of previous studies ................................ ...................... 59 2 . 2016 s demographic characteristics ................................ .................. 6 9 3. I nter correlations of key variables in the national level analysis ................................ ............. 80 4. R esults of SEM regression analysis about the relationship among personal factor s , risk perception, and behavioral responses related to climate change ................................ .............. 8 3 5. R esults of G SEM regression analysis about the relationship among personal factor s , risk perception, and eac h behavioral response associ ated with climate change .............................. 8 9 6 . 2015 s demographic characteristics ................................ ................ 10 4 7 . C orrelation matrix of key variables in the local level analysis ................................ ............... 11 4 8 . R esults of SEM regression analyses about the relationship among personal factor s , risk perception, and behavioral responses related to drought ................................ ........................ 11 6 9 . perception, and behavioral responses in the mitigation and adaptation contexts ................... 1 40 10 . Possible reason s for the insignificant relationship between experiences and risk perception ................................ ................................ ................................ ................................ ................. 14 3 11. Summary of the survey to members of the KWRA ................................ .............................. 1 5 9 12. Descriptions about the relationship between climate change experiences and climate change risk perception ................................ ................................ ................................ ....................... 1 60 13 . Summary of three empirical analyses ................................ ................................ ................... 1 71

PAGE 13

xiii LIST OF FIGURES FIGURE 1 . general system s model of human adjustment to natural hazards ....................... 41 2 . Rogers ................................ ................................ ........... 42 3. Grothmann and Patt model of private proactive adaptation to climate change .......... 4 4 4. Research model change ................................ ................................ ................................ ................................ ....... 6 0 5 . Pathways in the national level research model ................................ ................................ ......... 7 6 6 . Results of four long term climate change experiences ................................ ............................. 7 8 7 . Results ................................ .. 7 9 8 . Significant pathways in the national level research model ................................ ....................... 8 4 9 . Pathways in the local level research model ................................ ................................ ............ 1 11 10 . Results of seven water saving efforts than usual ................................ ................................ .. 11 3 11 . Significant pathways in the daily action and device utilization models ............................... 11 7 12 . Amount of damage caused by typhoons and heavy rains in South Korea (From 2002 to 2015) ................................ ................................ ................................ ................................ ................. 14 5 13 . Changes in the relationship between short term experiences and risk perception related to clim ate change ................................ ................................ ................................ ...................... 14 6

PAGE 14

xiv LIST OF ABBREVIATIONS ADF: Asymptotically Distribution Free CAT: Climate Action Tracker CFI: Comparative Fit Index COP: Conference of the Parties GHG : Green House Gas GSEM: Generalized Structural Equation Modeling ICT: Information and Communication Technologies INDCs: Intended Nationally Determined Contributions IPCC: Intergovernmental Panel on Climate Change K water: Korea Water Resources Corporation KWRA: Korea Water Resources Association MOLIT: Ministry of Infrastructure and Transport NASA: National Aeronautics and Space Administration OECD : Organization for Economic Cooperation and Development PMT: Protection Motivation Theory RCP : Repre sentative Concentration Pathway RMSEA: Root Mean Square Error of Approximation SEM: Structural Equation Modeling SRMR : Standardized Root Mean Squared Residual TLI: Tucker Lewis Index UN: United Nations UNFCCC: United Nations Framework Convention on Climate Change

PAGE 15

1 CHAPTER I INTRODUCTION Climate change is not a new issue. As early as 1896, the Swedish chemist S vante Arrhenius explored the relationship between atmospheric c arbon dioxide and air temperature to explain the existence of ice ages ( Arrhenius, 1896) . In 1964, social ecologist Murray Bookchin warned of the probability of anthropogenic climate change appearing and its potential adverse impacts (Bookchin, 1964). However, this issue failed to bring heightened attention from the general public and the inter national community for a long time. Over two decades later , the Intergovernmental Panel on Clim ate Change (IPCC) , which is the international organization for providing a scientific basis related to climate change to government s at all level s , was set up in 1988 (IPCC, 2017). Since then , the IPCC has continually warn ed of the risks of climate change through t he IPCC Assessment Reports ( IPCC, 2014a). A s a result, it has been highlighted as a major issue around the world. Global collaboration to address climate change has remained sluggish due to complicated political (e.g., numerous stakeholders and the absence of legitimate and coercive global organizations), economic (e.g., enormous costs to mitigate and adapt to climate change), and social (e.g., each circumstances. (Bodansky, 2012; Hagen, 2015; Klein, Schipper, & Dessai, 2005; Nyong, 2009; Stern, 2006) . Also, technical advancements linked to the reduction of Green h ouse Gas (GHG) emissions and the use of renewable energy are still insufficient and have not provided momentum to global collaboration. In 2015, the international community seemed to bear fruitful results by establishing the first global action plan to curb climat e change. This Paris C limate Agreement was endorsed by one hundred and ninety five countries ( European Commission,

PAGE 16

2 2017) , and the United States (U.S.) and China fulfillment ( Bodansky, 2016 ). A recent challenge to the agreement came early in 2017 when President Donald Trump declared that the U.S . would withdraw from the agreement ( Ellis, 2017). The prospect of international climate policy was dampened by this event ( Shear, 2017) . E ven though climate change is regarded as a global level risk by both the scientific and non scientific worlds , the national, local , and individual levels typically bear the costs of most impacts ( Altschuler & Brownlee, 2016; Mendelsohn, Dinar, & Williams, 2006) . To e ffectively alleviate the adverse effects of climate change, the world needs not only efforts at the global level, but also those at the national, local, and ind ividual levels. However, it is very difficult for both national and local policymakers to design and carry out effective and efficient climate change policies due to severa l challenges, such as the uncertainty around its predict ed trajectorie s and impacts (Stern, 2006; Tompkins & Adger, 2005 ) , juggling mitigation and adaptation strategies ( Biesbroek, Swart, & Van der Knaap, 2009 ; Kane & Yohe, 2000) , and the relatively low priority of climate change among national issues (Dunlap & Saad, 2001; Leiserowitz, 2005) . In particular , policymake rs in countries that suffer from a lack of resources (e.g., funds, experts, and technologies) are facing much more serious conditions . Fortunately, m any existing studies (e.g., Clayton et al ., 2015; Semenza et al ., 2008; Shove, 2010) advise that not only behavioral responses can be alternatives for alleviating the adverse impacts of climate change. Further , t he more behavioral responses to address climate change that individuals take, the less of a burden policymakers will feel in shaping and implementing climate policies. Thus, r evealing how to inspire individuals to more actively participate in coping action s against climate change might be no simple task, but be very meaningful. This dissertation investigates

PAGE 17

3 Recently, many researchers have paid attention to the question, what motivate s individuals to take behavioral respons es to curb climate change? Most of them focus on personal factors (e.g., experience and knowledge ) and climate change risk perception. It is true that p ublic risk perception has been spotlighted by scholars in that it can inform policymakers of public concerns 1998; Pidgeon & Gregory, 2004). 1 A t the same time, however, some previous studies (e.g., Battaglini et al Spence et al ., 2011 ) have noticed a different role of public risk perception , which is as a trigger of behavioral responses to address risk s . They note that higher levels of perceived climate change risk can lead to changes in personal actions. Also, personal factors are expected to influence behavioral responses to cope with climate change via risk perception or directly . Unfortunately , previous empirical research projects do not provide enough evidence about the process behavioral responses with respect to mitigating climate change and adapting to different types of climate change impacts (e.g., drought and flood risks). Most existing studies focus on the simple relationship between personal factors and climate change risk perception (e.g., Brody et al ., 2008; Lowe et al ., 2006; Slovic et al ., 2004) or between climate change risk perception and behavioral responses to address climate change (e.g., Battaglini et al ., 2009; Joshi et al ., 2013; Roco et al ., 2015) . Although a f ew r esearch projects have investigated the relationship among personal factors, climate change risk perception , and behavioral responses against climate change , they are usually based on specific industry contexts, such as agriculture ( Niles, Lubell, & Haden, 2013 ), or local level data (e.g., flood and 1 Public (risk) perception has been studied in diverse fields (e.g., health, technology, and environ ment) because policy making is affected by what the public perceives (Bostrom et al ., 1994; Leiserowitz, 2006; Ruddell et al ., 2012 ). In the climate change field, public risk perception can play a significant role in both shaping climate policies ( Brody et al ., 2008; Liu, Smith, & Safi, 2014) and generating public support for those policies ( Leiserowitz, 2005) .

PAGE 18

4 drought prone areas) ( Whitmarsh, 2008) . The findings of these studies are not necessarily applicable to behavioral responses of the general public at the national level. Also, w hile a number of studies (e.g., Lindell & Hwang, 2008; Takao et al ., 2004) explored risk perception and self protective behaviors by individuals against flood, research projects investigating personal risk perception and behavioral responses against drought are very rare (e.g., Deng, Wang, & Yousefpour, 2017) . To fill the gap in the literature and making linked to climate policy , t his dissertation tries to answer the following research question: How do personal factors (experiences, knowledge, and policy assessment) and risk perception related to climate change influence behavioral responses to address climate change? Obviously, t he aim of this study is not to explain all of the potential factors that drive individual responses to cope with climate change , but to shed light on this theoretical relationship and in doing so better articulate the role of risk percept ion in experience, know ledge, and policy assessments of climate cha nge . T his question is answered based on the case of South Korea, which is experiencing the rapid increase in GHG emissions in return for economic growth over the past decades. More specifically, the answer is derived from three different contexts : a) climate change risk perception and personal behaviors to mitigate climate change among individuals at the national level ; b) drought risk perception and personal behaviors to adapt to drought among individuals at the local level ; and c) significant issues affecting the relationship among personal factors, risk perception, and behavioral responses related to climate change as perceived by experts . Prior to empirical analyses, five hypotheses, including the positive correlation between risk perception and behavioral responses related to climate change, are proposed based on literature reviews. The works of Kates (1971), Rogers (1983), and Grothmann a nd Patt (2005) provide a theoretical

PAGE 19

5 foundation for developing the hypotheses. Two data sets , provided by the Korea Water Resources Corporation (K water) , are utilized to conduct quantitative analyses at the national and local levels . K water collected n at ional level data by a Web based survey ( 2016, N = 1,500) in South Korea, and collected local level data f rom a face to face survey ( 2015, N = 515) in n orth west area of South Ch ungcheong Province, South Korea. 2 R egression analysis, specifically structural equation m odel ing , is employed to evaluate the data . T o reinforce the findings of the two cross sectional analyses, this study conducts q ualitative analyse s based on an expert survey (2018, N = 98) and interviews (2018, N = 13 ) collected from the members of the Korea Water Resource Association, the K water institute , and other academic organizations . 3 By providing empirical evidence about the relationship among personal factors, risk pe rception, and behavioral responses to address climate change at the national and local level s along with expert knowledge, academically, this study contribute s to improving knowledge of behavioral responses against climate change as well as advancing the methodology of unveiling that relationship by using mixed met hods research. In addition, i t is meaningful in that this study will be an initial step toward revealing how climate change risk perception serves as a link between personal factors and behavioral response s related to climate change in South Korea. Practically, this dissertation inform s policymakers of critical clues as to how the general public voluntarily takes action to address climate change . P olicymakers both in South Korea and in countries with si milar socio economic conditions can benefit the most from this study 2 Two www.ope and conducted the Web based survey in 2016 and the face to face survey in 2015 respectively, by signing survey contracts with K water. 3 Those data include expert knowledge not only about the causal relationship among personal factors, risk perception, and behavioral responses to address climate change, but also about significant issues surrounding the relationship among them (e.g., obstac les which hinder the relationship and governmental policies to encourage

PAGE 20

6 This dissertation consists of seven chapter s , inc luding this introduction chapter . Two l iterature reviews . The second chapter covers the background literature explaining the fundamentals of climate change and behavioral responses against climate change r isk. It helps understand the obstacles of global collabora tion to address climate change, , and the features of climate change . The third chapter presents existing theories and empirical studies dealing with the relation ship among personal factors, risk perception, and behavioral responses related to climate change . A research question and hypotheses are provided in this chapter. Also, a research model based on the hypotheses is set up here. Three e mpirical analyses . The fourth and fifth chapters quantitatively test the hypotheses at the national and local levels , respectively . While t he former focuses on climate change risk perception and behavioral responses to mitigate climate change, the latter pays attention to drought risk perception and behavioral responses to adapt to drought. Both chapters provide information of research context , research method s , and the results of desc ri ptive and regression analyse s. These are followed by several policy implications and future research direction s , along with potential limitations with methodology . The sixth chapter qualitatively The key point here is to fill the gaps between the hypotheses and the findings of the two quantitative analyses. The components of this chapter are similar to those of the previous two chapters. Synthesis and conclusion . The seventh chapter briefly summarizes the content of the main chapter s (from II to VI) , including the results of the three empirical analyses , and evaluate s the attainment of study objectives . Finally , it presents useful lessons for policymakers and researchers respectively, and ends with a short epilogue.

PAGE 21

7 CHAPTER II BACKGROUND Fundamentals of Climate Change Despite the efforts of the Intergovernmental Panel on Climate Change and the international community, the outlook of future international climate policy remains bleak. Beginning of Anthropogenic Climate Change Since the existence of humankind, the interaction between human and environment has ensued in direct or indirect ways. During this interaction process, human activities often caused a disturbance of natural cycles that resulted in human misery ( Dankelman, 2010 ). Severe floods and desertification caused by deforestation of mountains (i.e., the overuse of natural resources) in human history might be good cases in point ( Dankelman, 2010 ). Climate change brought by global warming is somewhat similar to these di sasters in that human activities, called Greenh ouse Gas (GHG) emissions , are responsible for it. 4 However, climate change should not be treated equally as national or local disasters (e.g., earthquakes, wildfires, and hurricanes) . This is because it can trigger social, economic, and environmental damages (e.g., extreme weather event s, food shortage, biodiversity loss, and sea level rise) not o nly on a national or local level , but also on a global level ( Bellard et al . , 2012; Fischer et al ., 2005; Pearce et al ., 1996). More than a century ago, before the use of the term climate change, the Swedish chemist Svante Arrhenius investigated how atmospheric carbon dioxide is connected with air temperature to account for the existence of ice ages ( Arrhenius, 1896) . In 1964, the possibility of catastrophic climate change generated by human activities was cast by social ecologist Murray 4 increases in the concentration of the greenhouse gas (e.g., carbo n dioxide, methane, nitrous oxide, and chlorofluorocarbo ). Global warming triggered by this phenomenon leads to unpredictable climate change ( Houghton, Callander, & Varney, 1992; Nordhaus, 1991).

PAGE 22

8 Bookchin. 5 How ever, this issue raised sporadically by scholars did not receive much attention from the general public and the international community for a long time. Emergence of a Global Climate Change Organization development that meets the needs of the present without compromising the ability of future generat ions to meet World Commission on Environment and Development, 1987, p. 8), was introduced by ( Giddens, 2011) . The next year, a significant global effort for combating climate change first appeared. The Intergovern mental Panel on Climate Change (IPCC) was established jointly by the Wor ld M eteorological Organization (WMO) and the Unit ed Na tions Environment Program (UNEP) ( IPCC, 2014a) . The purposes of this intergovernmental organization are both to assess the causes and impacts of climate change and to come up with more effective measures for addressing this challenge. The findings of the an IPCC Assessment Report ( IPCC, 2014a). Some resea rchers have identified problems with IPCC studies , such as a lack of systematic empirical research ( Vasileiadou, Heimeriks, & Petersen, 2011), the increased mediatization and politicization ( Berkhout, 2010), and a monopoly on the provision of international climate policy recommendation s ( Tol, 2011) . D espite such criticisms papers have been rated as the most scientifically authoritative reports on climate change and have provided a framework for international negotiations related to climate change ( Bolin, 2007; Choi, 2016; IPCC, 2014b; Zillman, 2007 ). 5 Bookchin (1964) and oil) add 600 million tons of carbon dioxide to the by intercepting heat radiated from the earth, will lead to more destructive storm patterns and eventually to melting of the polar ice caps, rising sea

PAGE 23

9 Recently, the fifth assessment report points out that the main cause of global warming is GHG emissions from human activities, and human and natural systems are facing the greatest crisis due to climate change (IPCC, 2014b). According to the findings of the report, the average global temperature has risen by 0.89 degrees Celsius degrees Celsius ) for the past 112 Arctic glaciers will continue to recede and Antarctic glaciers have also declined in recent years per year. In other words, climate change is getting faster. Human caused GHG emissions are affected by several factors such as an economic activity, energy use, population size, technology, and climate policy . T he fifth assessment report provides the standard set of future climate change scenarios , called the Representative Concentration Pathways (RCPs) ( IPCC, 2014b). Simply, the R CPs refer to time dependent projections of atmospheric gree (Moss et al ., 2008, p. 5 ) . The f our RCPs described in the fifth assessment report consist of o ne stringent mitigation (RCP2.6) , two intermediate (RCP4.5 and 6.0) , and o ne high GHG emission (RCP8.5 ) scenarios . 6 According to IPCC (2014b), i f the international community does not make additional efforts to control GHG emissions, the pathway will be located between the RCP6.0 and RCP8.5. Given the RCP8. 5 , the global average temperature and sea level at the en expected to rise 3.7 degrees Celsius and 63 centimeters, respectively. H owever, if substantial net 6 Each RCP presents a quantitative description of climate change pollutant (e.g., CO 2 and CH 4 ) emissions and concentrations as well as radiative forcing in 2100 (IPCC, 2014b). Radiative forcing, measured in watts per square metre ( W/m 2 ( Bjørnæs, 2013, p. 2) . The numbers of RCPs mean their radiative forcings by 2100 (see Bjørnæs (2013), Moss et al . (2008), or Wayne (2013) for more information).

PAGE 24

10 negative GHG emissions that follow the RCP2.6 are realized by the international community , the rise of the global average temperature and sea level at the end of this century can be constrained by 1.0 degrees Celsius and 40 centimeters, res pectively. Finally, the report warns that i f GHG emissions continue (i.e., the efforts of GHG reduction fail), and the global mean temperature exceeds 2 degrees Celsius above pre industrial temperatures , damages due to climate change, such as the increase and the ri se of the death toll in extreme weather events , will most likely spread across the planet during and beyond the 21st century (IPCC, 2014b) . The IPCC has highlighted climate change as a universal issue, and its a ssessment reports have presented the major risks of climate change, which can have an irreversible impact on humankind ( Ford et al ., 2016; Zillman, 2007) . Also, t he IPCC has contributed to creating robust consensus that not only can effective global actions significantly reduce the levels of climate change risks, but they also lead to sustainable development ( Choi, 2016). It is true that the IPCC has shown science based leadership in the climate change fiel d. However, the IPCC is not a real actor on the climate change stage, but an assistant. It cannot solely combat climate change without activ e players , such as national and local governments and the general public . Global Efforts for Addressing Climate Change Since the 1990s, the international community has made efforts and attempts to combat climate change. However, the world is in a precarious situation and needs more efforts. In 1992, the efforts of the international community to address climate change began in earnest as the United Nat ions Conference on Environment and Development (UNCED) adopted the United Nations Framewor k Convention on Climate Change (UNFCCC) in Rio, Brazil ( Aldy & Stavins, 2012) . This convention was based on the scientific results of climate change risk coming from

PAGE 25

11 the IPCC and aimed at stabilizing the concentration of GHG in the atmosphere. The principles of the UNFCCC were ibilities and respective capabilities (CBD RC) these serve d as the core guideline of climate change negotiations ( Brunnée & Streck, 2013 , p. 589 ) . With the UNFCCC as a beginning, the efforts of the international community to prevent severe damages triggered by climate change gradually have expanded . Since 1995, parties to the UNFCCC have held the Conference of the Parties (COP) once a year for an international agreement on GHG emission control ( Bodansky, 2001) adopted the Kyoto Protocol, which specified global actions for responding to climate chang e. The Kyoto Protocol prescribes bind ing GHG reduction goals for thirty eight advanced countries (e.g., France, Germany, and Japan) which is to cut GHG emissions by 5.2 percent compared to the 1990 level durin Meyerson, 1998) . However, doubts about the effectiveness of the Kyoto Protocol ari se among the targeted nations. This is because a lthough some countries, such as China, South Korea, and India, have increased GHG emissions along wit h the economic growth, they are not obliged to cut GHG emissions ( Choi, 2016; Helm, 2008 ) . The commitment of the Kyoto Protocol weakens further as the U. S. , one of the most powerful countries and la rgest GHG emitter, r efuses to ratify the Kyoto Protocol ( Freedman & Jaggi, 2011) . To prepare for the post 2012 climate system after the K 13, he ld in Bali, Indonesia, adopted the Bali Roadmap. in Copenhagen, Denmark, attempted to draw up a new agreement, which forces both advanced and the rest of countries to participate in the post 2012 climate system, but this conference ended without big progress ( Park er et al ., 2012) . In 2010, t he efforts of the international community to

PAGE 26

12 C ountries participating in the conference agreed to put joint efforts into re alizing a long term goal (i.e., limiting the average temperature rise to within 2 degrees Celsius above pre industrial levels) through adopting a new agreement (Liu, 2011). Namely, the necessity of a paradigm shift to a low carbon society was recognized by the international (Hourcade & Shukla, 2015) . In 2013, 19, held in Warsaw, Poland, produced an agreement that each country submits a GHG reduction plan for 2020 and beyond under the nally Determined Contributions ( IND Cs) UNFCCC, 2017). I n 2015 , adopted the Paris A g reement. It stressed the principle of equality and common but differential responsibilities and acknowledged different national circumstances (Choi, 2016). Through this a greement, the GHG emission reduction system, which was centered on some advanced countries at the Kyoto Protocol, was expanded to all countries. This agreement endorsed by one hundred and ninety five countries initiated a global action plan to preven t drastic climate change through curbing global warming ( European Commission, 2017) . From a long term perspective, it aimed to limit the increase in global average temperature to well below 2 degrees Celsius above pre industrial levels (Rogelj et al ., 2016). To this end, each country should establish its GHG reduction plan along with adaptation plan, and its implementation performance will be checked by the United Nations (UN) (Savaresi, 2016). As of 2016, one hundred and eight nine countries submitted INDC reports related to the GHG reduction plan to the UN, which account for more than ninety seven percent of global GHG emissions ( Kitous et al . , 2016) . D espite the future prospect of climate change at the global level does not seem bright . According to the Climate Action Tracker (CAT) , glo bal

PAGE 27

13 temperatures will likely increase 3.2 degrees Celsius above pre industrial levels in 2100 even if major countries abide by their INDC reports (CAT, 2017 a ). 7 Additionally , t to withdraw fr om the Paris Agreement significantly deteriorates t hat figure compared to 2.8 degrees Celsius estimated in 2016 (CAT, 2017a) . Given that many states and cities in the U.S. have vowed to continue uph olding the a greement (Plumer, 2017; Regan, 2017) , the real impact of exit from the deal is likely to be smaller Nevertheless , it might be true that the withdrawal intention of the U.S. from the Paris Agreement has cast a dark cloud over global collaboration to address climate change ( Shear, 2017) . Conditions Surrounding Global Collaboration Complex c haracteristics of climate change . Regarding climate change , the Earth, more specifically the atmosphere, seems to have the features of public goods (i.e., non excludable and non rivalrous). 8 9 These goods (e.g., national security and fresh air) are usually accompanied by free rider problems and insufficient provision ( Bodansky, 2012) . A free rider refers to a person who g ains a benefit without paying all or part of its cost ( Gauthier, 1986) . Also, negative 7 The Climate Action Tracker , a scientific analysis, is provided jointly by three research organizations such as the Climate Analytics (http://climateanalytics.org/), the Ecofys (https://www.ecofys.com/), and the NewClimate Institute (https://newclimate.org/). It evaluates the efforts of thirty two nations (e.g., the E.U., China, and South Korea), which cover around eighty percent of global emissions, based on their INDC reports and provides expected global warming consequences over the 21st century ( CAT, 2017b) . 8 Public goods have two characteristics: Non excludable and non rivalrous ( Bodansky, 2012; Bueno de Mesquita, 2016; Kaul, Grunberg, & Stern, 1999 ). Bueno de Mesquita, 2016, p. 105). Recently, the extended concept of public goods , called global public goods , has been studied by many researchers (e.g., Barrett, 2007; Kaul et al ., 2003; Kaul, Grunberg, & Stern, 1999). This concept is simply defined as excludable, non Kaul, Grunberg, & Stern, 1999, p. 452). 9 Ostrom et al . (1999) describe cl imate change as a new challenge of global commons. According to Ostrom, Gardner, and Walker (1994), goods are categorized into four types (i.e., public goods, common pool resources, toll goods, and private goods), based on the degrees of exclusion ficulty of excluding individuals from and subtractability ommon pool resources have difficult exclusion and high subtractability (Ostrom, Gardner, & Walker, 1994). Given the quality of the atmosphere which can be depleted, the perspective of Ostrom et al . (1999) is also right.

PAGE 28

14 externalities can be triggered by the excessive use of public goods. 10 However , climate change caused by GHG emissions is actually more perplexing and insurmountable than it seems . W hile s ome countries emit a greater amount of GHG with economic activities into the atmosphere and l losses via global warming , the larger GHG emitters do not compensate them for the losses. Moreover , GHG emissions are likely to be over supplied . If no one pays the costs (i.e., the reducti on of GHG emissions), a ll countries in one ecosystem , the Earth, will bear immense expenses (i.e., damages from climate change) . T hose negative impacts can even affect future generations. Looking closer , climate change accompanied by negative impacts across the world and generations should be categorized Kaul, Grunberg, & Stern, 1999, p. 456). 11 Unfortunately , negotiations between the party triggering global warming and the party influenced by global warming are hindered by significant challenges such as difficulties in meas uring damages and transaction costs (Gruber, 2013; Weimer & Vining, 2011). T here is no gl obal organization (or nation) which has legitimacy or coercion related to GHG emissions . 12 Further, p oor countries are much more vulnerable to damages of climate change than rich countries (Nyong, 2009) . The more climate change becomes serious, the more citizens of poor countries will be under the triple torture of 10 (e.g., air and noise pollution s) are more problematic. 11 Barrett (1999) also regards Ozone depletion and climate change as global public bads. Similar to public goods, public bads are non excludable and non rivalrous ( Kolstad, 2000) . However, while public goods provide positive utility to consumers and are typically under supplied, public bads bring disutility to them and are over supplied ( Kaul, Grunberg, & Stern, 1999) . 12 Legitimacy means that its members are predisposed to accept the actions of the organization, whether they agree wi th those actions or not ( Bickers & Williams, 2001). Coercion refers to the method of persuading its members to do actions by using force such as imposing fines, taking property, and arresting ( Bickers & Williams, 2001) . Even though there are several global organizations associated with GHG emissions, such as the IPCC and UNEP, all of them solely depend on agreements ratified by countries. Some of the most powerful countries militarily and economically (e.g., the U.S. and China) are the main emitters of GHG, and global agreements are vulnerable to

PAGE 29

15 abnormal weather conditions, the lack of food and quality water, and social disturbances and economic instability. 13 Overall, t h e se characteristics, along with the following unsettled leadership and insufficient technological advancement, are retarding global collaboration for responding to climate change. Unsettled c limate change l eadership . Many scholars consider leadership as a core determinant of success or failure in coping with transnational issues (Young, 1991). 14 As mentioned above, poor countries, in which a majority of the population derives their livelihood from agriculture or fishery effects (Nyong, 2009) . A lso, i f the international community does not reduce GHG emissions due to national self interest, damages of climate change are passed on to future generations in the forms of natural disasters, conflicts of generations, and a larger split between the haves and have nots. Climate change is not just a political or economic problem. It is deeply connected to ethics and requires the leadership of global leading countries . 15 Both the U.S. and China are two of the most powerful countries in various fields as well as the second and first largest emitters of GHG in the world, respectively ( Christoff, 2016). In 2001, Paul G. economy is so large, its diplomatic influence so great and its contributions to environmental 13 The economies of poor countries can easily crash because their industries based on agriculture and fishery are deeply influenced by weather conditions ( Fischer et al ., 2005) . Even tho ugh poor countries have contributed little to the intensification of GHG emissions in the atmosphere, they will go through more serious torment than rich countries due to weather disaster prone industries and the lack of capacity and funding ( Fischer et al ., 2005; Mendelsohn, Dinar, & Williams, 2006; Nyong, 2009) . 14 Considering the work of Northouse (2016), leadership consists of four components (i.e., a process, influence, influences a group of defined as a process whereby a country (or countries) affects a group of countries to reduce global GHG emissions and minimize the impact of climate change. 15 Previous studies state that ethical leadership can positively affect organizations through reducing counterproductive and unethical behaviors (Mayer et al judgement (Trevino et al ., 1999), and fostering dedication, altruism, and better work willingness of followers (Brown, Trevino, & Harrison, 2005).

PAGE 30

16 problems so extensive ... the United States must be part of international solutions to However, in the same year, the Bush administration refused to ratify the Kyoto Protocol ( Oberthür & Roche Kelly, 2008). This event trigg ered the schism within the international community. About s ixteen years later, the Trump administration once again shocks many countries across the world by announcing the U.S. intention to withdraw from the Paris Agreement ( Ellis, 2017; Shear, 2017 ) . In t he case of China, even though it has been recognized as one of the leaders in the climate change field by many countries ( Parker et al ., 2012), China has taken a passive stance and seems unwilling to sh ow strong leadership toward climate change ( Harris, 2017; Heggelund, 2007; Lewis, 2008) . I n 1999, David A. Heenan and Warren Bennis introduced co leadership to the climate change field. From their perspective, c limate change is a real global issue involved with many tasks and stakeholders, unstable circumstances, and severe damages, and one leader cannot be responsible for this huge and complicated issue. Heenan and Bennis (1999) stress that co leadership should not only perm eate all institutions as sociated with climate change (from the global to local levels), but it also shares values and aspiration to curtail the negative impacts of global warming. Based on their work, Nhamo (2009) provides a co leadership model for climate change governance (i.e. , the leadership vehicle). This model consists of seven leadership zones: mitigation, adaptation, policy and legal frameworks, technology, financing, social dynamics, Nhamo, 2009, p. 477) . While the co leaders hip model can be utilized at various levels ranging from the local to global levels, there is no clear guideline for who mitigation leaders should be, who te chnology leaders should be, etc. I t requires leaders to find their zones by themselves, and this pr ocess seems very obscure and subjective. In sum, while leadership can play a pivotal role in enhancing global collaboration to

PAGE 31

17 combat climate change , it is still immature. G lobal leading countries still do not show an example of pursuing the ethical norms for the world in the climate change field. Co leadership does not seem practical due to the absence of an explicit guideline. Karlsson et al . (2011) write, It is no easy task to provide effective leadership on such a multifaceted issue which involves high p. 104). Insuffici ent t ech nological a dvancement . Increasing efforts to prevent GHG emissions generate increasing economic costs . T hese huge economic costs have made many countries reluctant to actively participate in global efforts on climate change (Osberghaus et al ., 2010 ) . To address this cost prob lem, the world should make ceaseless efforts to develop long term and sustainabl e technologies , instead of depending on short term and end of pipe measures which trigger marginal cost surges ( Jaeger et al ., 1997 ) . 16 17 Such a technological breakthrough cannot be realized immediately, and the longer the international community waits in inventing sustainable technologies, the higher the cost barrier will be ( Jaeger et al ., 1997) . In 1997, Jaeger and colleagues advised that the European Union ( E.U . ) break through the cost declaring ambitious targets for short This advice may not be only for the E.U., but be for all developed nations. Actually, many countries have made constant efforts to enhance technologies and experiences related to climate change. For instance, i n 2007, the E.U. set the target of 20% renewable energy by 2020, and many of its member states, such as 16 Long term and sustainable technologies to curb GHG emissions cannot be predicted in detail. But, considering the work of Jaeger et al . (1997), technologies which can improve energy efficiency, increase the production of accelerate ecological innovation might be part of them. 17 End of pipe solutions refer to the inhibition of pollution emissions by utilizing additional measures without curtailing resource use ( Frondel, Horbach, & Rennings, 2007) or the treatment of the im pact on the environment triggered by human activities without adjusting the production process itself ( Hammar & Löfgren, 2010) .

PAGE 32

18 Denmark, Germany, and the Netherlands, have been speeding up the development of clean energy technologies ( Lewis & Wiser, 2007; Jacobsson & Lauber, 2006; Lund & Mathiesen, 2009). 18 Yet despite efforts, fossil fuel based industries are still prevalent in the world, and high level techno logies and huge costs linked to climate change mitigation and adaptation strategies are a burden to many technically and economically weak countries. Consequently , the development of inexpensive and easy to apply technologies related to climate change can strengthen global collaboration and provides the bedrock of transformation to a low carbon society. But, like leadership, technological advancement is still insufficient to make a breakthrough in enhancing global struggle with climate change. Necessity of more comprehensive efforts . According to Meijerink and Stiller (2013), The evidence is clear that climate change is occurring and affecting all nations ( Brecher, 2015). Abnormal weather conditions (e.g., extreme ly high temperatures, floods, and droughts), glaciers melting, and sea level rise are evident today. More seri ously, it is getting worse ( IPCC, 2014b ). This is because past and current global efforts, especially the reduction of GHG emissions, are not sufficient to address perilous climate change (Osberghaus et al , 2010). IPCC (2014c) warns that if the rise of the global average temperature exceeds 2 degrees Celsius above pre industrial temperatures , it will fall into a vicious cycle of raising the temperature, and any human efforts to suppress that will become meaningless. No further fe nce sitting can be allowed, and g enuine accountability and swif t c ountermeasures are needed . P aradoxically, the characteristics of climate change as a global public bad inform the international community of both the necessities 18 Even though many advanced countries (e.g., the U.S., China, Canada, and Japan) turned away from the Kyoto Protocol ( Odeku, Mav eneka, & Konanani, 2013) , the E.U. remained faithful to its roles as a global leader and has played a significant role in saving it ( Oberthür & Roche Kelly, 2008).

PAGE 33

19 and the difficulties of global collabor ation to address climate change. L eadership and technological advancement do not still create favorable conditions for global effort s to cope with climate change. Under this tough circumstance , to ad dress climate change at the global level , both developed and developing countries should make more efforts to enhance global collaboration and to raise their reduction target of GHG emissions . In addition , c onsidering the fragile global collaboration and the above mentioned unstable conditions that surround it , this dissertation argues that efforts not only at the national and sub national levels, but also at the non governmental level (especially individual level) are indispensable . Government and Individual s Responses to Climate Change To cope with climate change, policymakers must try not only to design and implement effective climate policies , but also t o activate voluntary behavioral responses. Disaster The identification of climate related disasters by policymakers, whether they are natural origin or anthropogenic, is one of the key points in shaping response policies . Baum, Fleming , and Davidson ( 1983 ) compare the characteristics of natural disasters (e.g., floods , hurricanes, and earthquakes) and those of technological catastrophes (e.g., the collapse of dam s and the leakage of toxic waste dumps and radiation) . They state that while both of them are sudden, powerful, destructive , and hard to predict , there are two significant differences between them. First, natural disasters are uncontroll able and bring short lived i mpacts, while technological catastrophes ) and pose long lasting threats ( Baum, Fleming , & Davidson , 1983 , p. 346) . Second, recognizing the guilty party of natural d isasters is difficult, whereas the guilty party of technological catastrophes can be more easily identified . T his study assumes that while man made disaster is a broader concept

PAGE 34

20 than technological catastrophes, their characteristics are almost identical. 19 In the case s of natural disasters which are uncontrollable , have short lived impacts , and responsible agents are difficult to identify , policy makers may aim to minimize the adverse consequences of these disas ters . In terms of man made disasters which are controllable, have long lasting threats , and have responsible agents that are more easily identified , policy makers may try to reduce the cause and probability of those disaster s through controlling the responsible parties. T he complex identity of climate change does not allow policy makers to simply select on e of these dichotomous strategies . The scientific community overwhelmingly has agreed that anthropogenic global climate change is occurring (Oreskes, 2004). The report also indicates that GHG emissions from human activities are the ma in cause of global warming ( IPCC, 2014b). W hen considering only the cause of climate change, it can be categorized as a man made disaster and seems controllable . However, climate change is accompanied by natural disasters, commonly called extreme we ather events such as floods, droughts , and heat waves . The pattern of extreme weather and natural hazard events are regarded as a key signature of climate change by scientists ( Mirza, 2003; Scheraga & Grambsch, 1998) . Swim et al . (2009 ) also regard Further, even climate change is controllable, it takes a great deal o f time. In short , climate change has the features of both a man made disaster and a natural disaster . C limate change policies must deal with both the causes and consequences of this anthr . 20 19 Park (2011) views man such as poor Also, Wiegman and Gutteling (1995) which compar e natural disasters with man made risks echo the conclusion of Baum, Fleming , and Davidson (1983). 20 Since the early 20th century, natural disasters triggered by human activities have been investigated by researcher projects (e.g., Bellegarde Smith, 2011; Simpson, 1986; Ulrich, 1935). These studies mostly focus on earthquakes.

PAGE 35

21 Difficulties in Government Response s to Climate Change Unfortunately, e stablishing and im plementing effective policies to address climate change is no simple matter for policy makers due to the following three problems . Uncertainty . T he great uncertainty surrounds climate change policy. 21 Kane and Shogren (2000) characterize climate change policy as complex and contentious . This is because the path of future technology, estimates of net damages, and adaptability of social and ec onomic systems Kane & Shogren , 2000, p. 92) . Tompkins and Adger (2005) also note that long time frames, u nsure climate change impacts, and unpredictable social and economic futures make it chal lenging to create climate policies . Policy ma ker s have to design policies, all ocate resources wisely to serve th e poli cies, and convince the public to support their decision s u nder uncertainty . Balancing responses . K eeping a balance between two response options confuse policy makers . Climate change policies consist of mitigat ion and adaptation ( Füssel, & Klein, 2006; Osberghaus, et al ., 2010; Pielke, 1998; Stern, 2006 ) . Simply, m itigation refers to actions and Tompkins & Adger, 2005. p. 563). Typically, mitigation bring s benefits to all climate sensitive systems at the global levels an d its lead time is decades, whereas adaptation provides benefits to selected systems at the local or regional levels and its lead time is from immediate to decades ( Füssel & Klein, 2006) . T he formulation of mitigation and adaptation policies has been largely independent ( Füssel & Klein, 2006) . Regarding t he structure of the 21 R egarding uncertainty, Deser et al . (2012) also view it as a main challenge in planning climate policies.

PAGE 36

22 IPCC, Working Group II and Working Group III deals with adapt at ion and mitigation, respectively ( IPCC, 2017) . According to Wilbanks et al . (2003), the perception that the reinforcem ent of adaptation policies can weaken the requirement for mitigation policies causes an adaptation and mitigation dichotomy . T hey state that this dichotomy is particularly prevalent where th e interests toward mitigation are not strong and the costs of adaptation are considered affordable. Tompkins and Adger (2005) also note that governments make decisions about climate change mitigation and adaptation in different policy domains (e.g., energy sector / non energy sector) . However, many studies (e.g., Biesbroek, Swart, & Van der Knaap, 2009; Bosello, Carraro, & De Cian, 2010; Jones et al ., 2007; Kane & Shogren, 2000 ; Laukkonen et al ., 2009 ) highlight that the apparent separation of mitigation and adaptation (or trade offs between them) might result in ineffective climate policies. Namely , t he optimal mix of the two options leads to more risk reduction with less costs. Kane and Yohe ( 2000) write , E ach pursued in a vacuum can easily lead to increased costs of policies and programs with no attendant reduction in climate Overall , p olicy makers must effectively and efficiently react t o climate change , but the difficulty lies in seeking a good balance be tween m itigation and adapta t ion . Prioritization . C limate change is likely to be low on the list of national policy priorities . Leiserowitz (2005) reports that while the environment and climate change have gained a high level of both awareness and concern from the American public, it has not been perceived as an imminent or a high priority danger by them. In a 2000 Gallup poll, the environment came in 16th the most significant problems confronting the nation today. Among thirteen environmental issues, global warming ranked 12th (Dunlap & Saad, 2001). This phenomenon that the public does not regard climate change as a priority concern is not only confined to the U.S. In 2001, the English public viewed many environmental issues, s uch as

PAGE 37

23 hazardous waste disposal and water and air po llution , as more concerning problems than climate change ( Department for Envir onment, Food and Rural Affairs , 2002). In 2009, global warming and climate change were ranked 13th among sixteen issues for the government (e.g., economy, unemployment , and crime ) by the English public ( Department for Environment, Food and Rural Affairs , 2009). Bord, ny environmental issues are less pressing concerns for the public than economic and social issues. L ow priority of climate change may cause policy of attention to climate policy and finally lead to belated responses. F s perspective, uncertainty, standing at the crossroad of mitigation and adaptation, and low policy priority associated with climate change act as barriers to set up optimal climate policies. I n reality, it might be challenging for policy makers , especially in technically and economically weak countries, to perfectly overcome these obstacles by themselves. Thus , along with efforts to solve those issues , additional efforts to promote p ublic participation in addressing climate change are necessary . The role of individuals in the climate should not be underestimated. Actually, individual and households , along with particular industries or large organizations, are major source s of climate change. For instance, in the U.S., the amount of energy consumed by individual and households at homes and in nonbusiness travel account s for around thirty eight percent of total energy ( Gardner & Stern, 2008) . Difficulties in Individual Response s to Climate Change Thankfully, responses to address climate change can have positive effect s on climate policies . Previous studies (e.g., Kane & Shogren, 2000; Kane & Yohe, 2000) report that personal adaptive responses should be considered during decision making related to public investment in mitigation. T his dissertation further behavioral responses to cope with climate change are influential factor s on both mitigation and

PAGE 38

24 adaptatio n policies, especially where the costs of national scale mitigation and adaptation are considered insurmountable. The more behavioral responses to cope with climate change individuals take, the less of a burden policymakers will feel in shaping climate policies and allocating limited resources to the policies . However, when focusing just on personal benefits and costs brought by behavioral responses , they are less likely to occur automatically . Recall, g overnment s generally have two options in addressing climate change (i.e., mitigation and adaptation ) and implement both mit igation and adaptation policies. 22 Meanwhile, i ndividuals can have the third option: N o response . 23 They can overlook climate change and depend on efforts to address climate change brought by global organizatio ns or governments . This third . M itigation response . GHG emission s trigge ring climate change are derived from fossil fuel energy use, agriculture, and deforestation ( Stern, 2006) . The first source accounts for around two thirds of GHG emissions ( Stern, 2006) . G overnment policies , such as the transformation from fossil fuel energy to renewable energy and regulation on energy intensive industries , are indispensable to reduce GHG emissions. However, given that energy is typically used by industries and individuals, behavioral adjustments also are needed to mitigate climate change . Semenza et al . ( 2008 ) view both government action s and cooperation from consumers as necessary things to mitigate global climate change. 22 Given the work of Smit et al . (2000), mitigation policy is social economic systems in 23 Actual ly, no response is not the exclusive option of individuals. Governments also can use this option against intention to withdraw from the Paris Agre ement), no response to mitigation , they cannot fully avoid the responsibility of adaption related to extreme weather events (i.e., key figures of climate change), such as preventing flood, securing water for drought, and strengthening a response system for heat waves. Overall, no response to climate change is not a realistic option for governments.

PAGE 39

25 I energy saving activities can vary such as recycling , r educing the amount of waste, using less electric facilities, driving less and using public transportat ion , and purchasing ener gy efficient car s (or electric vehicle s ) and facilities ( Bockarjova & Steg, 2014; Masud et al ., 2016; Rainear & Christensen, 2017; Whitmarsh , 2009 a ). But , voluntarily taking these actions to mitigate climate cha nge is not a simple task . F rom a short term perspective, benefits and costs are i ntermingled in human decision making . For instance , in the case of recycling, individuals should prepare spaces for recycling and non recycling materials separately and invest their time to recycle materials. These action s for recycling are accompanied by tangible costs (e.g., spaces and time) and intangible benefits (e. g., intrinsic satisfaction s linked to values) . 24 I n contrast, i n terms of using less electric facilities, individuals can obtain tangible financial benefit s ( e.g. , saving money on their electric ity bill s ), whereas they might accept intangible costs (e.g., inconvenience ) . 25 Regarding driving less and using public transportation, while individuals save their money (e.g. , gas and parking fees) and keep their health , the trade off is that they may spend more physical energy (e.g., more walking) and time. Namely, it is uncertain whether benefits minus direct and indirect costs brought by their behavioral response for responding to climate change , is positive or negative. While those who achieve positive net utility are likely to participate in behavioral responses to mitigate climate change , those who achieve negative net utility are not (i.e., selecting no response option). F rom a long term perspective, a s sows and an M itigating climate change has a potential to bring benefit s to 24 Many studies report that values (or worldviews), such as altruism and environmentalism, are influential factors on pro environmental behaviors ( Barr, 2007; Cla rk, Kotchen, & Moore, 2003; Corbett, 2005) or coping actions against climate change ( Akerlof et al ., 2013; Kahan et al ., 2012 ) (see p. 63 for more information about pro environmental behaviors and coping actions against climate change). 25 Some of them may get intangible benefits, too (intrinsic satisfaction based on values).

PAGE 40

26 world through reducing impacts of climate change. However, even though mitigation actions are conducted by current generations, potential benefit s are expected to be enjoyed by next generations. Recall, the lead time of mitigation is decades. Whi le those who value benefits for next genera tion s are likely to participate in those actions to mitigate climate change, those who do not value that for next generation s are not (i.e., choosing no response option). A daptation response . As mentioned earlier, a nthropogenic c limate change brings severe extreme weather events , posing a great threat to humankind and ecosystems, such as flood and drought . E xtreme weather event s themselves are not climate change but natural disaster s . 26 However, anthropogenic global warming make s extreme weather events more violent and volatile. Thus , extreme weather and natural hazard events getting more intense are viewed as parts of climate change ( Mirza, 2003; Scheraga & Grambsch, 1998) . Among them , f l ood and drought are the most common events across the world , and adaptation efforts against these disasters have been made at the national, local, and individual levels. While both flood and drought are related to water (quan tity) management, they are quite different from each other. Flood . According to Berz et al ( p. 458). F lood , usually led by extreme precipitation, cause s extensive damage not only to human life and properties , but also to ecosystems. 27 Briefly , it is a matter of uncontrollably excessive water in certain areas . From an individual s perspective, their adaptation responses toward flood , such as 26 The World Meteorological Organization (2016) defines an extreme weather as weather variable above (or below) a threshold value near t (p. 4). 27 an overflow or inundation that comes from a river or other body of water and causes or threatens and events that deliver more precipitation to a drainage basin than ( U.S. Geological Survey, 2017) . River flood, flash flood, and storm surge are generally considered as the main types of floods ( Barredo, 2007; Berz et al ., 2001; Jonkman, 2005 ).

PAGE 41

27 setting up sandbags in front of the house, purchasing flood protection devices (e.g., pumps), and moving expensive furnishings to the second floor ( Zaalberg et al ., 2009; Grothmann & Reusswig, 2006), can bring direct benefits (e.g., minimization of property losses and loss of life) to themselves . Numerous research projects (e.g., Lindell & Hwang, 2008; Miceli, Sotgiu, & Settanni, 2008; Takao et al ., 2004; Thieken et al , 2007 ; Zaleskiewicz, Piskorz, & Borkowska, 2002 ) have investigated self protective behavior s by individuals against flood . M any of them identify that individuals get involved with behavi oral responses to address flood based on perceived flood proba bilities , perceived risk to life and property, and others ( Bubeck, Botzen, & Aerts, 2012 ). Namely , there is a high possibility that individuals voluntarily ta ke action against flood for obtaining personal benefits , and these activities can help policymakers ease the burdens of climate adaptation policy. Grothmann and Reusswig (2006) , which highl ight the impacts of adap tive behaviors toward flood , note that r protective activities at flood prone urban areas can contribute to the prevention of monetary flood damage by eighty percent and the reduction of the need of public risk management. D rought . Due to the conflicting concepts held by diverse academic fields, drought, Lloyd Hughes, 2014, p. 607), has no unanimous definition ( Dracup, 1991; Dracup, Lee, & Paulson, 1980; Palmer, 1965 ; Wilhite & Glantz, 1985 ) . 28 Instead, the simple definitions of Palmer (1965) and Sheffield and Wood (2012) are useful for understand ing this event . Palmer (1965) defines drought as 28 Drought is usually categorized into three ( Dai, 2011; Dracup, Lee, & Paulson, 1980) or four ( Thurow & Taylor, 1999; Wilhite & Glantz, 1985) types: a) meteorological drought w normal Dai, 2011, p. 45) ; b) agricultural drought a average precipitation, intense but less frequent rain events, or above normal evaporation, all of which lead to reduced crop Dai, 2011, p. 45) ; c) hydrological drought a during which streamflows are Linsley, Kohler, & Paulhus, 1975, p. 374) ; and d) socioeconomic drought societ Mehran, Mazdiyasni, & AghaKouchak, 2015, p. 7520).

PAGE 42

28 Sheffield and Wood (2012) describe it as . In brief , drought can be viewed as a water shortage due to less precipitation than normal . Drought has several characteristics. From a general approach based on the work of Wilhite (2000) , f irst, the existence of drought itself is unclear due to the absence of a precise and objective definition of drought. Second, it is uncertain when drought starts and ends because of gradual accumulation and long term persistence of its effects. 29 Third, a wider range of geographical areas suffer s from drought than other natural disasters (e .g., flood s and hurricane s ), and the quantification of the loss and necessary relief works is challenging . From a policy approach ba sed on the work of Kim (2016) , first, governments are likely to miss the timing to have a sense of crisis because of creeping impact s . T hey expect a water shortage to be solved through one or two rain events. Second, when people or industries suffer from the restriction or interruption of the water supply for a long time belated responses, great social and economic d isruption can happen. Third, d espite the adverse effects of drought on local societies, drought is underestimated among natural disasters because it hardly bring s infrastructure collapsing and casualties. 30 Fourth, the severity of drought is determined not only by natural factor s (e.g., precipitation), but also by social factors (e.g., dam operation, wat er infrastructure, and water consumption pattern). Therefore, there can be a great deal of conflict regarding who is responsible for drought . Given these characteristic s of drought, i t seems more difficult for individuals to detect and address drought than flood. Also, u nlike flood, behavioral responses against 29 Zamani, Gorgievski Duijvesteijn, and Zarafshani (2006) categorize drought into ent slow evolvement, uncertainty and ambiguity, belated perception, and an unclear low point as the features of the slow onset disasters. 30 However, given the work of Peduzzi et al . (2003), severe drought in highly vulnerable nations (mostly poor countries) can bring numerous casualties.

PAGE 43

29 drought (e.g., saving water) generally bring social benefits rather than personal benefits. W hile behavior al responses to cope with flood influence personal life and properties (e.g. , house and furniture ) , those against drought affect water resource s (e.g., dams, reservoirs , and rivers), regarded as common pool resources and increase not personal resistance but social resistance to drought. 31 Further, w hether or not an individual take s water saving action s , he or she can have an equal opportunity of water consumption with others during the period of drought . Namely, individuals are more likely to be reluctant to take action to address drought compared to flood . In brief , i helping out government climate policies and addressing climate change . However, the process of responses based merely on personal benefits and costs triggered by the responses is not likely to be seamless and reliable . Advantages of Behavioral Responses to Climate Change Although the ac tive role of governments is more imperative due to unstable global collaboration associated with climate change, policymakers are also facing real issues in establishing and implementing climate policies, such as uncertainties, keeping a balance between mitigation and adaptation, low policy priority, and a lack of resources which is the most serious problem in develo ping countries. M any research projects have suggested policy frameworks or specific strategies related to climate change mitigation (Bosetti, Carraro, & Tavoni, 2009; Lutsey & Sperling, 2008) and adaptation ( Biesbroek et al ., 2010; Lemieux et al ., 2014; Lim et al ., 2004) to help policymakers . A ll of those studies are valuable to policymakers to manage climate policies. Especially, Hallegatte (2009) introduces several significant and practical considerations 31 Ostrom, Gardner, and Walker (1994) present two key points of common appropriators or limiting appropriation rights

PAGE 44

30 to policymakers when handling climate policies: a) t he production of benefits , whether or not ); c) the ); d) the utilization of ins ( 244) . Surprisingly, individ behavioral responses related to climate change mitigation and adaptation largely meet these options. 32 Consequently, c urbing the adverse impact of climate change requires not only Semenza et al ., 2008; Shove, 2010). Clayton et al . (2015) stress the importance of individual behaviors: Climate change cannot be effectively addressed without the exploration of its impact on human well being, public perception, and influential factors on mitigation and adaptation behaviors. Semenza et al . (2008) Opportunity for immediate action lies in increasing voluntary mitigation by individuals engaging in more sustainable, low Given that i ndividua behavioral efforts for responding to climate change can support climate policies and also have the above mentioned advantages, exploring how to responses against climate change (i.e., finding a more robust behavioral responses instead of depending on the relationship between personal benefits and costs from the responses ) and informing policymakers of the findings are significant assignments of the academic world . T b ehav ioral responses to address climate change, media ted by risk perception, in the next chapter. 32 health, and increasing satisfaction associated with the protection of environment (of course, these actions are accompanied by some costs such as inconvenience and time). Those actions also bring societies benefits, such as saving social resources, regardless of the existence of climate change. Furthermore, their activities are flexible and can be linked to institutional and financial tools.

PAGE 45

31 C HAPTER III THEORY Risk Perception D riving Behavioral Response s Despite the strong potential of risk perception in activating responses to climate change, empirical evidence surrou nding this relationship is still in sufficient. Climate Change Risk In light of previous research projects (e.g., IPCC, 2001; Team, Pachauri, & Reisinger, 2007; Whitmarsh, 2009b), climate change is a change in the typical or average weather caused by (direct and indirect) human activities or natural origin . 33 However, w hat inspires human responses to climate change is not climate change itself, but the perceived risk triggered by cl imate change. 34 Climate change risk can be described as the probability of exposure to perilous consequences on the health or living conditions of humans and ecosystems brought by anthropogenic climate change (or conditions led by pr evious climate change), considering existing studies (e.g., Böhm & Pfister, 2001; Jones, 2001; Pachauri et al ., 2014; UN, 1992). 35 Recall, temperature will prompt several risks such as a decrease in food production, an increase in the risk of extinction of land and freshwater species, and land loss due to coastal flooding (IPCC, 33 Weath er and climate are distinct terms. While weather refers to the short term conditions (changes) of the atmosphere in terms of temperature, precipitation, wind, etc., climate refers to average weather conditions over a long period of time (i.e., a statistica l phenomenon) (Lombardi & Sinatra, 2012; National Aeronautics and Space Administration (NASA), 2011; Rebetez, 1996; Weber, 2010; Whitmarsh, 2008). 34 Tompkins & Adger, 2005, p. 564). 35 The definition of risk varies depending on the research fields, such as health, economics, and environment ( Sjöberg, Moen, & Rundmo, 2004; Yates & Stone, 1992). Simply, the significance of those losses, and uncertainty associated with those (Yates & Stone, 1992, p. 23). Slovic, 2000, p. 220).

PAGE 46

32 2014b). These risks threatening hu man life and ecosystem are accompanied by untold economic damage. Estimates of damage range from US$ 140 billion to US$ 1.4 trillion (IPCC, 2014c). Also, climate change risk has been identified in diverse fields such as human health ( Haines et al ., 2006; McMichael, Woodruff, & Hales, 2006; Patz et al ., 2005) , food (Lobell et al ., 2008; McMichael et al ., 2007; Rosenzweig & Parry, 1994) , ecosystem s (Millar, Stephenson, & Stephens, 2007; Walther et al ., 2002; Winder & Schindler, 2004) , and water management ( Arnell, 2004; Christensen et al ., 2004; Vörösmarty et al ., 2000). Much scientific evidence indicate s that human human societies ( Klein, Schi pper, & Dessai, 2005; Whitmarsh, 2009a). Given the prevalence of climate change risk, this poi nts to the to what degree is the general public aware of (or familiar with According to Brechin and Bhandari (2011), at least sixty six percent of the respondents in all the twenty five countries surveyed (e.g., Germany, South Korea, and the U.S.) believe that the problem of global warming is either very serious or somewhat seri ous . 36 Based on the different cross national survey data, they note that more than sixty percent of the respondents in all the fifteen countries surveyed (e.g., China., India, and Japan) answer that either people are being substantially affected by climate change now or it will happen within twenty five years . 37 Hagen (2015) reports that more than half of the respondents in the eight countries surveyed (e.g., Germany, Spain, and the U.S.), except the Netherlands, are either highly concerned or concerned about the possible impacts of 36 More than fifty percent of the respondents in seventeen out of the twenty five countries answer that the problem of global warming is very serious. The question has a four point scale: very serious , somewhat serious , not too serious , and not a problem . These survey data are derived from the Pew Global Attitudes Project s tudy released in 2009 ( Brechin & Bhandari, 2011) . 37 Also, a considerable major of the respondents in all the fifteen countries indicate that climate change will have significant human and ecological impacts. The results are based on the data of the World Bank survey administered in 2009 ( Brechin & Bhandari, 2011). The findings of other comparative studies related to climate change perception, such as Kim (2011) and Leiserowitz (2007) , are in line with those of the above mentioned studies .

PAGE 47

33 climate change. Over sixty five percent of the respondents in all the nine countries believe that if global warming increase s substantially over the next twenty years, either serious negative consequences or moderate negative consequences will take place. 38 Over the past three decade s , the general public has been exposed to scientific evidence and academic findings crystallizing climate change risk through mass media, public campaigns, social communities, private communication networks (e.g., family and friends) , and others. It is likely to be difficult for the general public to exactly figure out what trajectory climate change follow s (will follow) and how much climate change negatively affect s (will affect) humanity and ecosystems (in the future) in detail. However, this study situation has increased . I ndividuals are increasingly aware that climate change is a serious risk , an on going risk , and a gradually intensifying risk . Role of Public Risk Perception Regarding the fundamental question about what goes on in individuals mind s and w hat thought process they usually go through when reacting to climate change, Swim et al . (2009 ) o ffer a critical clue. They write , (Swim et al ., 2009, p. 37). This dissertation argues that understanding risk perceptio n related to climate change is a prerequisite for reveal ing against climate change. perceived likelihood of negative consequences to oneself and society from one specific Based on existing studies (e.g., Böhm & 38 The survey ques tions have a five point scale, where 1 is not at all concerned (not likely at all to happen) and 5 is highly concerned ( serious negative consequences ). The data of cross Global Survey on Public Attitudes towards Clima Hagen , 2015, p. 68) .

PAGE 48

34 Pfister, 2001; Sjöberg, Moen, & Rundmo, 2004), this study more precise ly defi nes climate change risk perception as a subjective evaluation of the probability of clim ate change (or conditions produced by previous climate change) happening and the severity of and vulnerability to its adverse consequences . The role of risk perception is presented in the following paragraphs from two perspective s : A policy perspective and an individual s behavior perspective . From a policy perspective . As alluded to previously , the establishment and implementation of effective climate policies at the national and local levels is no simple task, and depending solely on scientific inf ormation might not be a desirable way to soften this situation . This is due to the difficulties of conveying complex scientific informa tion associated with global environmental change (e.g., climate change and stratospheric ozone depletion) to policymakers and of making usable information for decision making purposes (Pielke, 1994). Also, scientific information cannot clearly answer polic , 1994, p. 317). In other words , it is very hard for policymakers to distill practical significance of scientific information related to global environmental change, to generate policy alternatives, and to secure public support. Instead of scientific information, government policies rela ted to environmental risks are subject to public debate and risk p erception, which attract policy et al ., 1981; Johnson & Tversky, 1983). Public risk perception has widely been regarded as a significant factor in pushing or rest ricting political, economic, and social actions to cope with particular risks ( Leiserowitz, 2005). Many researchers also highlight that public risk perception can drive both technological and scientific risk assessments and relevant policies (Correia et al ., 1998; Perry, Lindell, & Tierney, 2001). Moreover , u nderstanding public risk percep tions is critical in developing effective policies to alleviate both human induced adverse impacts on the

PAGE 49

35 environment and human vulnerability to climate change impacts ( Ruddell et al ., 2012) . In short, p ublic risk perception play s a significant role in produc ing effective climate policies and inspiring public support. From an individual s behavior p erspective . Both nature and society scholars are fundament ally interested in the question: H ow do people perceive and respond to climate change? (Diggs, 1991). In the health field , in which investigations related to risks and behavioral responses are prevalent, estimating the seriousness of a risk (risk perception) , comparing th e benefits and costs of responses (coping appraisal) , and choosing the response maximizing expected outcomes (behavioral response) are common steps in the process of addressing the risk (Van der Pligt, 1996). Likewise, understanding how people perceive environmental risks is critical to predicting their responses to environmental issues (Fischhoff, 1985). I n the climate change field, previous research projects (e.g., Battaglini et al ., 2009; Bockarjova & Steg, 2014; Keshavarz & Karami, 2016; Niles, Lubel l, & Haden, 2013; Yu et al ., 2013) have investigated the relationship between climate change risk perception and individ of them verify that individual perceptions of climate change lead to their beha vioral responses to address climate change . 39 Overall, p ublic risk perception can play a critical role in addressing climate change both This study mentioned that individuals are becoming more aware of climate change risk i n the previous section . Nevertheless , it may not be easy for the general public to clearly perceive climate change risk 39 A few different point of views also exist. For instance, Mainieri et al . (1997) warn that public perception related to the environment may not perfectly correspond with behaviors. Based on survey data in the Athens area, Greece, Papoulis et al . (2015) report that while the majority of the respondents are concerned about climate change and agree with public actions, they remain unlikely to take individual action to cope with climate change. This is because the respondents believe that climate change should be solved by the efforts of state and industry, which they view as the main stakeholders, rather than those of individuals.

PAGE 50

36 due to the three major characteristics of global change: a) the long temporal scales, which hinder with everyday life; and c) its inherent abstraction ( Kearney, 1994) . Also, d ifficulties of transferring scienti fi c information are likely to hinder the general public from perceiving climate change risk ( Rebetez, 1996) . Th erefore, close attention should be paid to the formation of climate change risk perception as well as the process to address climate change , mediated by that risk perception. Gaps in Climate Change Risk Perception Literature Since the early 1970s, when the first Earth Day and the initial peak of the ecology movement emerged (Erskine, 1972), many research projects have investigated public perception regarding environmental issues. public has gained persistent concern in the U.S. The work of Dunlap and Scarce (1991) also supports the evidence of growing public concern toward environmental issues and public support for government actions to address those issues. Regarding individual behavioral responses, Baldassare and Katz (1992) unveil that those who perceive a serious environmental threat to their health and well being have more willingness to engage in eco friendly practices (e.g., recycling and less driving) based on survey data in the Orange C ounty, California . Since the 1990s, research projects related to climate change risk perception also have been increasing with many scientific warnings. The characteristics of the existing studies are as follows:

PAGE 51

37 First, initial studies, such as Read et al . (1994), Bostrom et al . (1994), and Berk and Schulman (1995), focus on how people think about climate change and how risk perception affects the related climate responses based on interview data in the U.S. Later, her (1999) explore the relationship between climate change risk perception and willingness both to support government actions (e.g., tax and regulation) and to take voluntary actions (e.g., driving less, avoiding the purchase of a gas guzzler, and using le ss air conditioning and heat). These topics have been consistently explored by many scholars (e.g., Berk & Fovell 1999; Brouwer, Brander, & Van Beukering, 2008; Veronesi et al ., 2014; Viscusi & Zeckhauser, 2006; Zahran et al ., 2006) . Most of them indicate that the majority of people are concerned about climate change, and risk perception has a positive correlation with willingness to pay. Second, since the 2000s, many studies have paid attention to the relationship between personal factors and climate chang e risk perception. These personal factors include past experiences ( Keller, Siegrist, & Gutscher, 2006; Slovic et al ., 2004) , physical vulnerability ( Brody et al ., 2008), knowledge (Shi et al ., 2016; van der Linden, 2015), political preferences (Carlton & Jacobson, 2013; McCright & Dunlap, 2011), and diverse socio demographic characteristics (e.g., gender, age, education, and income). However, this research neglects the linkages between personal factors and behavioral responses addressing climate change. Th ird, while research projects about the relationship between risk perception and behavioral responses have been growing recently, most are limited to certain fields, such as farming ( Battaglini et al ., 2009; Niles, Lubell, & Haden, 2013; Roco et al ., 2015), stock farming ( Joshi et al ., 2013; Opiyo et al ., 2016), fisheries ( Musinguzi et al ., 2016), and tourism (Saarinen

PAGE 52

38 & Tervo, 2006), which can be directly damaged by climate change. Unfortunately, the findings of these studies which are mostly relat ed to specific adaptation strategies of certain industries are not necessarily applicable to behavioral responses of the general public. Fourth, even though some studies (e.g., Deng, Wang, & Yousefpour, 2017; Spence et al ., 2011; Whitmarsh, 2008 ) have expl ored the causal relationship among personal factors, risk perception, and behavioral responses associated with climate change, these empirical research projects typically employ a single personal factor (e.g., experiences of flood, drought, or climate chan ge), focus on specific industry contexts (e.g., agriculture) and adaptive responses, or utilize the data derived from certain cities and flood and drought prone areas. Overall, numerous research projects have studied climate change risk perception. However, their findings do not provide enough evidence about the process of personal factors risk perception behavioral responses to mitigate climate change which can be generalized in the majority of people at the national level. Additionally, very few st udies (e.g., Deng, Wang, & to cope with drought, whereas a number of studies (e.g., Lindell & Hwang, 2008; Takao et al ., 2004; Thieken et al ., 2007; Zaleskiewi cz, Piskorz, & Borkowska, 2002) have explored risk perception and self protective behaviors by individuals against flood. Filling the Gaps in the Literature: Research Question Research q uestion . Climate change has already For alleviating the adverse effects of global change, human behaviors must be changed ( Kearney, 1994). This dissertation also asserts that what national level policymakers really need i s how to encourage people voluntarily participate in behavioral responses to mitigate climate change, and what local level policymakers actually need

PAGE 53

39 Therefore, making related to climate change, this study investigates the following research question: How do personal factors and risk perception related to climate change influence person al behaviors to address climate change? This research question is answered in three different contexts: a) climate change risk perception and personal behaviors to mitigate climate change among individuals at the national level; b) drought risk perception and personal behaviors to adapt to drought among individuals at the local level; and c) significant issues influencing the relationship among personal factors, risk perception, and behavioral responses r elated to climate change as perceived by experts. Thr ee p ersonal factors . Unlike physical scientists who approach risks from empirical climate observations and models, the risk perceptions of the public are being affected both by scientific and technical descriptions of risk and by a broader spectrum of factors such as psychological, social, and personal experiences ( Lorenzoni & Pidgeon, 2006 ; Slovic, 2000). As mentioned above, many personal factors (e.g., experiences and knowledge) have been utilized by previous research projects. This dissertation focus es on three personal factors: Personal experience , knowledge , and policy assessment . It is important to note that this study does not aim to understand all of the factors that can shape behavioral responses to address climate change. Inste ad, it aims to better understand the connections among these factors, risk perce ption, and behavioral responses and articulate the role of risk perception as a mediator. While this approach may not capture some of the antecedent variables that a re theoreticall y and behavioral responses to climate change, it allows this study to zoom into factors that are both recognized across different literatures as proximal to behavioral responses, and which have policy relevance (also .

PAGE 54

40 Integrated Research Model Higher r isk perception affected by p ersonal factors ( exp erience , knowledge, and policy assessment) related to climate change may intensify ind Behavioral Responses Previous research projects investigating th e causal relationship among personal factors, risk pe rception, and behavioral responses associated with climate change lack theoretical, conceptual , and methodological coherence. To enhance the theoretical foundation and to derive hypotheses and a research model , this dissertation introduces three works, Kates (1971), Rogers (1983), and Grothmann and Patt (2005). These three studies provide insights that suggest risk perception acts as a causal mechanism linking personal factors and behavioral responses to address the risk . First, Kates (1971) proposes a general system s model of human adjustment to natural hazards. identify key components of the human response system, to present the strengths and direction of the linkages among those components, and to simulate the processes derived from the components and linkages. Simply, he posits that hazard experiences influence risk perception , and risk perception leads to human responses. Figure 1 Kates, 1971, p. 447) . s, and these risk perceptions lead to adjustment strategies. However, i t does not specifically describe a process of forming risk perception. 40 40 While some natural hazard related studies (e.g., Hanson, Vitek, & Hanson, 1979; Mitchell, 1974; Pulwarty & Riebsame, 1997) have intro duced the work of Kates (1971), no empirical research project related to climate change employed

PAGE 55

41 Figure 1 . general system s model of human adjustment to natural hazards Second, Rogers (1975; 1983) introduces Protection Motivation Theory (PMT) . He presents the cognitive steps of , which include is an Rogers, 1983, p. 153). 41 Briefly, the purpose of PMT is to r eveal how a threat leads to coping actions (i.e., behavior changes) through protection motivation . 42 It has widely been applied in diverse fields, such as health ( Floyd, Prentice Dunn, & Rogers , 2000; Milne, Sheeran, & Orbell, 2000), technology ( Ifinedo, 2012; Vance, Siponen, & Pahnila, 2012) , and environment ( Bockarjova & Steg, 2014; Kim, Jeong, & Hwang, 2013; Rainear & Christensen, 2017) based on broader applicability . three parts: The acquisition of information, perceptual process, and responses. First, people obtain information from environmental and intrapersonal sources. Second, the acquisition of 41 Rogers (1983) writes these sentences associated with a fear appeal, based on the sermon of Edwards (1741), 42 how individuals process threats and select responses to

PAGE 56

42 information initiates the perceptual process. Throughout this process, people compare rewards with the threat severity and their vulnerability (i.e., the expectancy of being exposed to the threat), called threat appraisa l . They also compare self and response efficacy with response cost. This process is called coping appraisal . 43 Lastly, protection motivation is formed by the results of the two appraisals, and it leads to certain responses. Figure 2 lays out the schema of protection motivation theory ( Rogers , 1983, p. 168). PMT also depicts a consecutive process among influential factors, risk perception, and coping actions. Figure 2 . Rogers Some research projects applied PMT to natural hazards and climate change. In the case of natural hazards, Mulilis and Lippa (1990) utilized PMT to study whether a negative, threat inducing message Martin, Martin, and Kent (2009) investigated the relationship among direct experience with wildfire, risk perception, and the risk mitigation process based on the data derived from three 43 These appraisals can be described by the following question (1) Does an individual obtain rewards from the threat? (i.e., perceived rewards) (2) Does an individual believe that the threat is severe? (i.e., perceived severity) (3) Does an individual feel vulnerable? (i.e., percei ved vulnerability) (4) Does an individual believe that the adaptive response can be an effective way of avoiding or minimizing the threat? (i.e., response efficacy) (5) Is an individual confident in his or her capacity to accomplish the adaptive response? (i.e., perceived self efficacy) (6) Are the costs related to the adaptive response small and affordable? (i.e., perceived response costs) (Cismaru, Lavack, & Markewitch, 2009; Rogers, 1983).

PAGE 57

43 homeowner associations in the western U.S. Poussin, Botzen, and Aerts (2014) applied PMT to flood preparedness in three flood prone regions in France. Overall, the findings of thes e studies support . In terms of climate change, Cismaru et al . (2011) in vestigated a mix of eleven communication campaigns associated with climate change deriving from English ACT ON CO 2 K ingdom (U.K.) Ireland, Climate Change Household Action Campaign in Australia The study reports that most PMT variables that affect behavioral changes are utilized by all of the campaigns. Rainear and Christensen (2017) explored whether the five PMT constructs (e.g., severity, vulnerability, and response costs) have a relationship with intentions to engage in pro environmental behaviors (i.e., mitigating global warming). The findings of this study stro ngly match PMT. 44 Similarly, t he work of Kim, Jeong , and Hwang (2013) , which examine s behavioral intentions related to climate change mitigation among American and Korean students , reports that some components of PMT (e.g., perceived severity and response efficacy) are significant predictors of behavioral intentions to cope with climate change. Lastly, Grothmann and Patt (2005) develop a process m odel of private proactive adaptation to climate c hange PMT, This model has two major cognitive processes, called climate change risk appraisal and adaptation appraisal , ada ptation efficacy, self This model has several features compared with PMT. First, considering the cha racteristics of 44 While the work of Rainear and Christensen (2017) strongly verifies PMT, it has some limitations. Their work focuses on the stage of cognitive mediating processes. The authors do not utilize specific influential factors, such as experiences and knowledge, except socio demographic characteristics. Also, respondents are limited to a certain job (i.e., students enrolled in a large public university) and a certain area (i.e., a northeastern part of the U.S.).

PAGE 58

44 climate change which generally do not bring benefits, this model does not reflect intrinsic and extrinsic rewards in the risk appraisal process. Second, reliance on public adap ta tion, which negatively affect risk appraisal, is newly added to the model. 45 Third, an additional cognitive process, called avoidant maladaptation c onsisting of fatalism, denial, and wishful thinking, is placed between the appraisals and the adaptation intention (i.e. , 46 This model posits that influential factors, such as personal experiences, public adaptation, and social discourse, have a relationship with risk perception, and this perception formed by the two appraisals influence personal adaptive responses. Figure 3 describes Grothmann and Patt (2005) model ( Grothmann & Patt, 2005, p. 204). Figure 3 . Grothmann and Patt model of private proactive adaptation to climate change 45 Grothmann and P adaptation policies conducted by public agencies will make people less likely to take action to adapt to climate change. 46 Avoi dant maladaptation, also called maladaptive coping or nonprotective responses , has been operationalized by many previous studies ( e.g., Abraham et al ., 1994; Bubeck, Botzen, & Aerts, 2012; Fruin, Pratt, & Owen, 1992; Grothmann & Reusswig, 2006; McCrae, 1984). Rippetoe and Rogers (1987, p. 598) provide the definitions of the following core modes: a) avoidance a and d) hopelessness

PAGE 59

45 Research Modeling Approach It is true that models have the potential to analyze the causal relationship among personal factors, climate change risk perception, and behavioral respo nses to address climate change . However, this study argues that an adjusted research model, which is more practical and tailored to the research question and policymak , is necessary when considering the following three important points. First, diverse personal factors that can be linked to the establishment of climate policy should be included in the research model. an experience factor. 1983, p. 167). 47 ience, knowledge, and policy related factors. 48 While this research project focuses on key personal factors identif i ed by t he three work s (e.g., experience, knowledge, and policy assessment ) , these personal factors will be specified to provide multifaceted policy implication s based on the findings of previous Second, cognitive processes connecting personal factors and behavioral responses need to be simplified. S ome components in model , such as response efficacy, self efficacy, and response costs , play an important role in deciding adaptive behavioral responses . Generally, adaptive behavioral responses in certain 47 Also, cognitive mediating processes (p. 167). Knowledge related to risks may be transmitted during the process of verbal persuasion. But, this study does not view it as an explicit knowledge factor. 48 Actually, social discourse on climate change risk and adaption is not a personal factor, but it might affect personal knowledge related to climate change. Also, Grothmann and Patt (2005) p osit that personal knowledge affects behavioral responses via the adaptation appraisal.

PAGE 60

46 fields (e.g., agriculture and stock farming) can bring both significant costs and benefits directly to people who take action , and a significant level of capacity is required to accomplish those responses . For instance, when farmers suffering from climate change alter their crops, which are less vulnerable to climate change than current ones, they might consider these question s: a) how much can these crops endure climate change and minimize the losses caus ed by climate change ? (i.e., response efficacy); b) i s it possible for me to grow these crops? (i.e., self efficacy); and c) can I afford co sts triggered by changing the crop and d o the additional benefits derived from altered crops outweigh the costs? (i.e., response costs). However , i n the case of the general behavioral responses associated with climate change mitigation or adaptation (especially drought) , they typically do not pay high costs and do not receive direct and significant be nefits from their behavioral responses ( small benefits and costs are often mixed together) . H igh capacity is not necessary for completing behavioral responses. Also, m any mitigating activities (e.g., less driving, energy saving, and others) and their significant impacts on the reduction of GHG emissions are well known to the general public , and they are not difficult to practice . Namely, those efficacy related components research model. Besides, data from two surveys utilized in the following national and loca l level analyses unfortunately do not include the measurement of efficacy related components . Thus , the research model excludes the adaptation appraisal and just focus es on the climate change risk appraisal (i.e., c limate change risk perception). Third, the ways in which different personal factors influence behavioral responses to address climate change can vary. The works of Kates (1971), Rogers (1983), and Grothmann and Patt (2005), suggest that personal factors, such as experiences and knowledge, influence risk perception through a cognitive process, and risk perception a ffects behavioral responses .

PAGE 61

47 present a different approach. They do not regard risk perception as an independent cause of environmentally friendly or hostile behaviors, but an essential by product of environmental beliefs. Namely, posit t hat b ehavioral responses to address climate change are a partial function of climate change risk perception. Some studies pay attention to both a direct path (from pe rsonal factors to behavioral responses ) and an indirect path (from personal factors to behavioral responses via risk perception) . For example, Niles, Lubell, and Haden (2013) investigated the relationship among fa r factors, risk perception, and responses in the U.S. T hey identify both the impact of a direct path (e.g., from climate change experience to government program participation) and an indirect path (e.g., from climate change experience to government program participation via climate change risk perception). Zaalberg et al . (2009) explored the relationship among past flooding experiences, cognitive appraisals, and coping efforts in the Netherlands. They also veri fy the direct and indirect impacts through using mediated models predicting coping responses. Therefore, there is a possibility that responses to address climate change are functions of both personal factors and climate change risk perception. To clarify these functions, the researc h model should be designed to explore the dual path impacts, directly between personal fact ors and behaviors as well as indirec tly via ris k perception . Hypotheses and Research Model Hypotheses . Based on the identified theoretical foundation , research modeling approach, and the findings of existing empirical studies, five hypotheses are proposed. Experiences . Many studies (e.g., Keller, Siegrist, & Gutscher, 2006; Slovic et al ., 2004; Weinstein, 1989 ) note that risk perception is deeply influenced by past experiences . Regarding climate change, the perception and reaction of the public are largely based on their own

PAGE 62

48 experiences, Rebetez, 1996, p. 495). Spence et al . (2011) also stress the importance of the first to climate change (p. 46). These arguments are theoretically supported by the works of Kates (1971), Rogers (1983), and Grothmann and Patt (2005), which consider personal experiences as one of the key influential factors on risk perception and behavioral responses. Empirical evidence about the relationship between personal experiences and risk perception related to climate change has been provided by previous studies . This evidence is based on the experiences of certain e xtreme weather events, such as flood ( et al ., 2005 ; Siegrist, & Gutscher, 2006; Whitmarsh, 2008) and drought ( Dessai & Sims, 2010; Diggs, 1991), or the experiences of global warming (Akerlof et al ., 2013 ; Donner & McDaniels, 2013; Li, Johnson, & Zaval, 2011 ). For example, Diggs (1991) studied whether drought experience s affect based on a mail survey in t he western solidified by d rought experience s . et al . (2005) reveal that managers of community water systems who have experience of a flood emergency in the last five years show a higher flood risk perce ption than managers who do not by analyzing survey data in two eastern American states (South Carolina and Pennsylvania). The work of Akerlof et al . (2013) explored whether perceived personal experience s of global warming (e.g., changes in seasons, weather, and lake levels) have a relationship with risk perception in the They conclude that respondents who have more perceived personal experience of climate change present higher perceptions of local climate change risks. Based on long term climate and survey data, Donner and McDaniels (2013) in vestigated the impact of national temperature fluctuations

PAGE 63

49 on concern s about climate change in the U.S. The authors report that mean temperature anomalies are correlated with greater concern s about anthropogenic climate change . While these studies identify the positive correlation between personal experiences and climate change risk perception, Whitmarsh (2008) provides mixed results. He explored how knowledge, attitudes, risk perception , and behavioral responses related to climate change based on interview and survey data in the south of England. The study reports that the experience of floods makes little vic climate change differently. The number of s tudies on the relationship among personal experience, risk perce ption, and behavioral efforts related to clima te change has been Bard, and Fisher (1999) state that perception related to the bad consequences of climate change has a significant positive relationship with both voluntary actions to mitigate climate change and voting intentions on climate policy proposals. Lubell, Zahran, and Vedlitz (2007) hypothesize th at perceived climate change risk positively influences global warming activism (i.e., policy support, political participation, and environmental behaviors). Their empirical test based on telephone survey data in the U.S. successfully verify the hypothesi s. Based on survey data in China, Yu et al . (2013) investigated whether public concern about climate change and perception of the impact of climate change affect willingness to take individual actions to address climate change. While public concern about cli mate change has a significant positive relationship with behavioral intentions, perception of the impact of climate change shows mixed results. Spence et al . (2011) investigated the relationship among personal experience of flooding, concern over

PAGE 64

50 climate c hange, and willingness to save energy based on the data of the national survey across the U . K . in 2010. They reveal not only the positive correlation among them, but also the direct and indirect (via risk perception) impacts of flooding experience on energy saving behavioral intentions through utilizing path analysis. Meanwhile, based on structural equation model ing , Deng, Wang, and Yousefpour (2017) explored whether drought experience s influence two perceptions (i.e., climate change and saving water) and water saving behaviors in only b etween perception and adaptive behaviors . In the case of Niles, Lubell, and Haden (2013) , they ., water availability), climate change risk perception , and government pro gram participation through employing path analysis. While the past experience significantly affects the two variables, surprisingly, it has a negative relationship with the governmen t program participation. Personal experience itself does not ensure a positive relationship with risk perception. This is because each experience is based on the diverse conditions of severity and vulnerability and can bring different subjective appraisals about risks. While experiences of climate change with high severity and vulnerability can present a positive relationship with risk perception, climate change experiences with low severity and vulnerability can have no relation to risk percepti on and even can have a negative relationship. Also, because risk perception is likely to be shaped by other factors (e.g., core values, education, and gender), the impact of these experiences can be mitigated by these other factors. Howev er, previous st udi es have found that the impacts of climate change are getting more severe, and many nations and local areas a re

PAGE 65

51 highly vulnerable to those of climate change. T he majority of existing empirical studies identify positive relationsh ips between personal experiences and risk pe rception associated with climate change. To better understand the structural relationship among personal experiences, risk perception, and behavioral responses to address climate change, this study divides person al experiences related to climate change into two types: Short term experience and long term experience . 49 This approach reflects the definition of personal climate change experience described in Appendix A. The short term experience consists of the personal experiences of extreme weather events (i.e., drought and flood). The long term experience is composed of t he personal experiences of four climate change features (i.e., the rising of average temperature, the change in precipitation, the increase o f the danger of flood, and the increase of the danger of drought in the last ten years). While some research projects (e.g., Mase, Gramig, & Prokopy, 2017; Niles, Lubell, & Haden, 2013) view statements about climate change features (e.g., the global climate is changing and average global temperatures are increasing) as climate statements about climate change features based on a long term period (e.g., at least a decade) can be treat ed as experiences . 50 Based on the identified theoretical approaches ( i.e., Grothmann & Patt, 2005; Kates, 1971; Rogers, 1983 ) and much empirical evidence , this study posits that personal experiences (both the short term and long term experiences) related to climate change have a positive correlation 49 The words, short and long terms , are subjective, and there is no clear guideline about them. Considering the this dissertation views a long term as at least a decade. Füssel and Klein (2006) also note that the lead time of mitigation is decades. 50 Climate change, which is a long temporal, extensive spatial, and far from our daily environment phenomenon, cannot be easily experienced by individuals (Swim et al ., 2011). However, recent studies (e.g., Akerlof et al ., 2013; Howe et al ., 2013) indicate th at climate change in the local level has been fairly accurately detected by individuals. Therefore, this study views the long term experience as perceived experiences. Nevertheless, it might be true that s is tightly intertwined with beliefs.

PAGE 66

52 climate change risk perception. Additionally, given some studies ( e.g., Niles, Lubell, & Haden, 2013; Spence et al ., 2011; Zaalberg et al ., 2009 ) have identified the direct impact of past experiences on behavioral responses, along with the indirect impact via risk perception, this study also expects the existence of the direct and significant path impact between the short term experience and behavi oral responses to address climate change. T he first h yp othesis is proposed: Hypothesis 1 : The more experience with climate change people have , the higher level of climate change risk perception they will have (and in the cas e of the short term experience, the more actions to address climate change they will take ) . 51 Additionally, t o analyze the relationship among personal experiences, drought risk perception, and adaptive behaviors against drought, experience of water supply problems (e.g., limited water supply, low water pressure, etc.) during a drought period is utilized as a n experiential factor . Knowledge . The knowledge theory, which is the most widely held theory of risk people perceive technologies (and other things) to be dangerous Wildavsky & Dake, 1990, p.42). 52 PMT posits that cognitive mediating processes (i.e., shaping risk p erception) , which lead to coping actions , ca n be initiated by observational learning. s (2005) model considers know ledge as a significant factor influencing climate change risk perception. (1971) model also implicitly indicates that knowledge influences human adjustment s to cope with natural hazards. 51 Experience with climate change and climate change risk perception include experience with drought and drought risk perception, respectively. 52 Wildavsky and Dake (1990) investigated the relationship between several influential factors (e.g., knowledge, personality, political orientation, and cultural biases) and risk perception. They identify a minimal relationship between self rated knowledge and risk perception.

PAGE 67

53 Knowledge r elated factors have been used in a number of empirical research projects. There is no doubt that knowledge is a potential factor increasing climate change risk perception or behavioral responses to address climate change. For instance, Malka, Krosnick, and Lange r (2009) studied the relationship between knowledge and concern a bout global warming based on data from two survey s in the U.S. They conclude that more knowledge leads to greater concern among Democrats, Independents, and people trusting scientists to convey dependable information. In New Zealand, Milfont (2012) explored the relationship among knowledg e, risk perception, and perceived efficacy related to global warming and climate change by utilizing panel data (three time points from 2008 to 2009) . The author reports that concern s about global warming and climate change are significantly intensified by knowing more about them. 53 Lee et al . (2015) examined risk perception. By analyzing the data of 119 countries, which derived from the Gallup World Poll in 2007 2008 , they note that climate change risk perceptions are strongly predicted by the understanding of human caused climate change, especially in Latin America and Europe. In the same vein, b ased on online survey data of six countries ( e. g. , Canada, Chi na , the U.K., and the U.S.), Shi et al . (2016) investigated the relationship between three knowledge factors (i.e., physical characteristics, causes, and consequences) associated with climate change and concern about climate change. The research findings unveil that the knowledge factors, espe cially the causes of climate change, are important drivers of climate change risk perception. onnor, Bard, and Fisher (1999) use knowledge, measured by the level of understanding about the causes of climate change, as an independent variable when assessing 53 Also, Milfont (2012) stresses that higher concerns about global warming and climate change increase perceived efficacy and responsibility to help address them.

PAGE 68

54 voluntary actions to alleviate climate change and voting intentions about climate policy prop osals. The y find that knowledge has a significant posi tive relationship with both voluntary actions and voting intentions. Halady and Rao (2010) explored the linkage between awareness to climate change and behavioral changes amongst individual managers in India. They report that behavioral changes of individual managers are significantly influenced by awareness to climate change. In the case of Yu et al . (2013), they categorize the causes of climate change as a perception factor instead of a knowledge facto r. The results of their study are consistent wit h previous studies: The less that people perceive climate change as natural consequences, the more that people take action to cope with climate change. Meanwhile, Lubell, Zahran, and Vedlitz (2007) hypothesize global warming knowledge as a positive factor on global warming activism (i.e., policy support, political participation, and env ironmental behaviors). U nlike the perceived risk factor, which has a positive relationship with the three ac tivities, they find that global warming knowledge significantly and positively affects only one activity (i.e., policy support). Based on survey data from an Alpine region in Switzerland, Jaeger et al . (1993) also reveal that while the level of educational achievement is a significant predictor of climate relevant behaviors, knowledge about climatic risks is not. Cons idering the findings of both theoretical and empirical studies, knowledge , mostly associated with general featur es of climate change (e.g., causes and consequences), typically shows a positive correlation with climate change risk perception or coping actions against climate change. The catch here is that many empirical studies employing knowledge related factors (e.g., Kellstedt, Zahran, & Vedlitz, 2008; Malka, Krosnick, & Langer, 2009 ; Milfont, 2012 ; Shi et al ., 2016 ; Stevenson et al ., 2014) focus not on behavioral responses to address climate change, but on climate change risk perception. W hile some studies (e.g., Halady & Rao,

PAGE 69

55 2010; Lubell, Zahran, & Vedlitz, 2007; or, Bard, & Fisher, 1999) verify the statistically significant relationship between knowledge and behavioral responses to address climate chang e , they do not consider that climate change risk perception can be a significant medium for connecting them (i.e., the process of personal knowledge risk perception behavioral responses ) . 54 Also, in light of the identified theoretical foundation ( i.e., Grothmann & Patt, 2005; Kates, 1971; Rogers, 1983) , knowledge associated with climate change , especially general characteristics of climate change , should be regarded as an influential factor mediated by climate change risk perception on behavioral responses to address climate change . To investigate behavioral responses to mitigate and adapt climate change , t his study utilizes four general knowledge factors , which are commonly used by previous studies , and two specific knowledge factor s related to government policy. In the case of mitigation , three factors (i.e., a wareness of climate change, the understanding of the cause of climate change, and the familiarity with a government GHG reduction plan ) are used as knowledge related factors . I n terms of adaptation , the remaining three factors (i.e., awareness of water quantity condition s , the exposure to water saving information , and the familiarity with subsidy policy during a drought period ) are utilized. The application of policy related knowledge can give policymakers an opportunity to witness whether or not the knowledge diffusion related to climate policy behavioral responses to address climate change . However, n o research project provides clear evidence about t he impacts of climate policy related knowledge on risk perception or behavioral responses related to climate change. T his study asserts that the 54 Both Lubell, Zahran, and Vedlitz (2007) and perception and knowledge as influential factors on behavioral responses to address climate change independently. Their findings reveal that the impact of climate change risk perception is much stronger than that of knowledge on behavioral responses to cope with climate change.

PAGE 70

56 familiarity with government policies to address climate change will lead people to support the se policies through their behavioral responses to cope with climate change. 55 In sum mary , this research project views general knowledge of climate change and climate policy related knowledge as influential factor s related to climate risk perception and indivi As an exception , the impact of t he exposure to water saving information , one of the general knowledge related adaptation s , is investigated through both direct and indirect paths . It participation in a campaign, publicity, or education related to water saving. Usually, these events not only inculcate people with general knowledge related to drought and action guidelines t o save water, but also motivate them to ta ke action. In other words, unlike the other three kn owledge factors, this factor is likely to influence both climate change risk perception and Hypothesis 2 and 3 are proposed: Hypothesis 2: The more general knowledge associated with climate change people have , the higher level of climate change ris k perception they will have (and in the case of the impact of the exposure to water saving information , the more actions to address climate change they will take) . Hypothesis 3: The more policy related knowledge associated with climate change people have, the more actions to address climate change they will take. Policy assessment . Unlike personal experience and knowledge, policy assessment has received minimal attention from scholars . Adger et al . (2009) write Policies and non climatic drivers currently play perhaps an even more important role [than climatic drivers] in 55 According to Connor, Bard, and Fisher (1999), knowledge can directly influence environmental behaviors when providing cues for appropriate behavioral strategies. Knowledge related to government policies generally give the general public useful clues what it takes in or der to address certain risks.

PAGE 71

57 (p. 335). However, onl y a few empirical studies utilize assessment factors related to government policy. For example, in 1999, onnor, Bard, and Fisher used government helpfulness (i.e., the scale for government working well) as an votin g intentions to climate policy proposals (e.g., t ax and regulation). The findings of the study reveal that the government helpfulness and voting intentions have a significant positive relationship. The research framework of Yu et al . (2013) easured by the level of trust in the effectiveness of the government climate change. Based on survey data in China, they unv eil that the policy assess ment factor has a significant positive relationship with individual behaviors to curb climate change. Also, past experience with environmental policies, clima te change risk perception, and their response (i.e., intention about government program par ticipation). While prior research finds a significant negative relationship between past policy experience (positive policy assessment) and risk perception , and a significant positive relationship between risk perception and government program participation, the direct impact of past policy experience on government program participation is not detected. From an empirical perspective , 99) utilize a policy assessment factor, but the authors do not investigate the relationship between government helpfulness and voluntary actions to address climate change. W hile Yu et al . (2013) identify a significant relationship between a p olicy assessment related factor and behavioral responses to address climate change, t hey do not consider that climate change risk perception can act as a bridge between them . Lastly, Niles, Lubell, and Haden (2013) fail to verify the direct

PAGE 72

58 relationship between personal policy assessment and behavioral responses , whereas their findings validate the process of personal policy assessmen t shaping risk perception, which in turn shapes behavioral responses related to climate change. From a theoretica l perspective , Grothmann and Patt ( 2005) i nc reliance on public action process model of private proactive adaptation to climate change. Reliance on public actions has a negative against clima te change. Namely, they expect that individuals change will decrease if climate policies impleme nted by public agencies can successfully address climate change risk. Given the results of empirical research projects and the theoretical approach of Grothmann and Patt (2005), policy assessment associated with clim ate change can be regarded as a negative influential factor on climate change risk perception. Four policy assessment factors are utilized to explore the relationship with climate change risk perception. Concerning mitigation, drought policy assessment and flood policy assessment are emp loy ed. Regarding adaptation, water supply assessment and water infrastructure assessment are utilized. Hypothesis 4 is proposed: Hypothesis 4 the higher level of climate change risk percepti on they will have. Risk perception . Based on the identified theoretical approaches (i.e., Kates, 1971; Rogers, 1983; Grothman n & Patt, 2005) and the findings of numerous empirical studies , the fifth a nd last hypothesis is proposed: Hypothesis 5: The higher people perceive climate change risk, the more actions to address climate change they will take.

PAGE 73

59 Controls: Socio demographic characteristics . Many existing studies associated with climate change risk perception and behavioral responses to address climate change include socio demographic characteristics , such as age, gender, and education , in their research model . While some factors significantly affect risk perception and behavioral responses, other factors do not. Also, no fac tor always influence s the dependent variables. For example, Semenza et al . (2008) reveal that education and age have a significant relationship with behavioral changes to address climate change. In the study by , only gender significantly affects willingness to take voluntary actions to reduce GHG emissions, whereas gender, age, and education significantly influence voting intentions to climate policy proposals. Yu et al . (2013) report that none of their demographic fac tors (i.e., age, gender, and the length of residence) are significantly correlated with behavioral intentions to address climate change . Table 1 . Significant s ocio demographic factors of previous studies Authors Socio demographic Factors Affected Factors Bard, & Fisher (1999) Gender(Female, +)*** Behavioral intentions Age(+)*, Education(+)** Policy preferences Leiserowitz (2006) Gender(Female, +)*, Member of environmental groups(+)** Risk perception Political i deology ( **, Member of environmental groups(+)** Policy preferences Zahran et al . (2006) Gender(M Policy preferences Semenza et al . (2008) Behavioral responses Keshavarz & Karami (2016) Behavioral responses Rainear & Christensen (2017) Race(no White, +)*, Age(+)** Behavioral intentions * Significant at .05, ** significant at .01, *** significant at .001; all two tailed tests. Note 1. Behavioral intentions and responses mean willingness to take individual actions and self reported actions to mitigate (or adapt to ) climate change, respectively. 2. Policy preferences refer to the extent of supporting climate policies (e.g., taxes and regul ations) to curb climate change.

PAGE 74

60 Table 1 arrays the significant socio demographic factors of several studies related to climate change. When testing the five hypotheses, this study reflects five socio demographic factors in the research model. Considering the commonly used factors of previous research projects, gender, age, education, and income are selected as control variables. Lastly, the type of a residential area is added to explore the difference between urba n and rural areas. While some int er personal factors, such as political affiliations, could be relevant theoretically in explaining behavioral responses, the se factors are outside the scope of this study. Research Model . Consequently, a research model is provided in Figure 4 , based on climate change risk perception, three types of personal factors (i.e., experience, knowledge, and policy assessment), and five socio demographic factors (i.e., gender, age, ed ucation, income, and residential area). Additionally, key concepts to test the hypotheses are defined and operationalized in Appendix A. Figure 4 . Research model climate change

PAGE 75

61 Potential Issues in the Methodology Scope of the research . Many empirical studies have explored various influential factors on climate change risk perception, such as values or worldviews (e.g., egalitarianism, individualism, and fatalism) ( Akerlof et al ., 2013; Leiserowitz, 2006), media coverages (Leiserowitz, 2004; Lowe et al ., 2006), past experiences ( Spence et al ., 2011), and knowledge (Shi et al ., 2016). van der Linden (2015) states that four dimensions (i.e., cognitive knowledge, experiential affect and personal experience, socio cultural culture, values, worldviews, and social construction, and socio demographic gender, race, income, age, education, and political ideology) primarily affect climate change risk perception. While this dissertation cove rs van der Linden three dimensions (i.e., cognitive, experiential, and socio demographic), socio cultural dimension, such as values (deep core beliefs) or worldviews about climate change , is also likely to be associated with risk perce ption or behavioral responses including policy preference, especially based on research in the U.S. (e.g., Akerlof et al ., 2013; Kahan et al ., 2012; Leiserowitz, 2006; Smith & Leiserowitz, 2012) . 56 A ssessing socio cultural dimension is outside the scope of this study, which is a possible limitation of the research but a potentially important area for additional study . However, given the context of South Korea whi ch is a relatively smal l size, largely homogenous , and same cultural country, it may not be a critical problem, as deep core beliefs around climate change (e.g. , belief in climate change) do not vary as much in South Korea as in the U . S . , and values/ w orldviews positions could be more homogenous . 57 56 Values (deep core beliefs) or worldviews may even get intertwined with personal factors. For example, could be informed by their experience and knowledge. Nevertheless, this dissertation differentiates experience and knowledge from beliefs, and in line with much of the literature. 57 previous studies (e.g., Lo, 2013; van der Linden, 2015) as a potential source of influence on behavioral responses to climate change. The role level analysis.

PAGE 76

62 Additionally , from the scope of this dissertation has a practical advantage. Evidence about the impacts of the socio cultural factors derived from different contexts (e.g., cultures, countries, or societies) might not be practically useful information in shaping climate policies change within their own cultures, countries, or societies. Even if policymakers try to utilize the socio cultural or them to control those factors, and, even if possible, it may take a significant amount of time and effort. Instead, evidence about the impacts of three factors (i.e., experience, knowledge, and policy assessment) and risk perception can be relatively easy to apply information in any contexts when addressing climate change. Also, these factors Measure of b ehavioral r esponses . Some studies, such as (1999) , Yu et al . (2013), and Rainear and Christensen (2017), employ behaviora l intentions to responses . According to Ajzen and Fishbein (1980) , behavioral intentions have a correlation with actual behaviors . 58 Meanwhile, Grothmann and Pat t (2005) hold a different view: Even though people have intentions, they often do not translate their intentions into actual behaviors due to a lack of resources (e.g., time, mo ney, and staying power), which are not considered during the process of forming the intentions. Some others (e.g., Deng, Wang, & Yousefpour, 2017; Keshavarz & Karami, 2016; Semenza et al ., 2008) utilize self reported behavioral responses instead of behavioral intentions. Gifford, Kormos, and McIntyre (2011) warn of the discrepancies between self reported behaviors and actual behaviors brought 58 Ajzen and Fishbein (1980) present two important factors for the correlation between intention and behavior. First, 51). Second, be as close as possible to the behavio ral observation within proper behavioral criterion.

PAGE 77

63 by seve ral bias factors such as personal wish es toward strong pro environmental scores and social desirability. Namely, regarding the reliability of behavioral response data, t here can be a difference between self reported and observed behavioral responses. However , in reality, it is al responses at the national a nd local levels requires painstaking efforts, time, and money. A survey method has many advantages, such as providing more effective means of social description, acquiring many information from a single survey, and saving considerable time and money, compa red to observation ( Singleton & Straits, 2017). Overall, despite problems triggered by the bias factors, this study practically utilizes self reported behavioral response data based on the survey. Relationship with pro environmental action literature . Pro environmental behaviors, which behavior that consciously seeks to minimize the have been investigated by many previous studies (e.g., Da hlstrand, & Biel, 1997; Karp, 1996; Kollmuss & Agyeman, 2002; Sia, Hungerford, & Tomera, 1986; Ster n, 2000; Vining & Ebreo, 1990). Actually, to pro environmental behaviors (e.g., using less energy, eating vegetarian food, and saving water). However, while pro environmental action literature has generally focused on social norms, environmental values, and personal propensity as influential factors, this study zooms into the influence of three personal factors (experience, knowledge, and policy assessment) and risk perception on behavioral responses to mitigate long term climate change and adapt to severe drought. Despite some overlap, this research is worth exploring , and its findings could provide insights both climate change and pro environmental action literature s .

PAGE 78

64 CHAPTER IV NATIONAL LEVEL ANALYSIS To explore the relationship among personal factors, climate change risk perception, and behavioral responses to mitigate climate change, this research p roject analyzes survey data of individuals from across South Korea . Given the context of South Korea which has rapidly achieved economic growth over the past decades and has experienced the increase of GHG emissions, the findings of this analysis can be meaningful n ot only to South Korea, but potentially many countries with similar socio economic conditions in managing GHG emissions and shaping climate policies. Research Context s Economic D evelopment After the 1953), South Korea has been able to achieve rapid pment oriented growth plans, strong leadership, a highly educated and hardworking workforce, and large business groups chaebol ( Amsden, 1989, p. 9; Yoo & Lee, 1987). Its rapid economic growth, was accompanied by the advancement of heavy chemical industries and manufacturing ( Yoo & Lee, 1987, p. 95). Even though the South Korean economy f aced a financial crisis in December 1997, the Korean government promptly overcame that crisis in 2001 through painstaking efforts, such as downsizing the government, restructuring economic systems, privatizing several state owned enterprise s, and fostering information technology ( Kim & Moon, 2002). 59 Since the 2000s, South Korea has been regarded as one of the leading 59 The International Monetary Fund declared a liquidity crisis and asserted that South Korea was experiencing a ( Crotty & Lee, 2009, p. 149). In the process of overcoming this crisis, large numbers of businesses went bankrupt, and South Korea struggled against extensive national level economic restructuring and severe public sector reform (Ito, 2007).

PAGE 79

65 countries in the Inform ation Communication Technology (ICT) ( Larson & Park, 2014; Lee & U ngson, 2008 ). For instance, A ccording to the ICT Development Index released by the International Telecommunication Union , it ranked the 1st and 2nd out of the one hundred and seventy six countries in 2016 and 2017, respectively (International Telecommunication Union , 2017). This social and economic development helped yield an increase of GDP per capita, which rose from US$944 in 1960 to US$25,459 in 2016 (World Bank , 2018). 60 G reenh ouse Gas E mission s R apid economic growth is likely to trigger the increase of GHG emissions. As of 2014, Korea ranked 4th (around 568 million tonnes) among thirty five Organization for Economic Cooperation and Development ( OECD ) member states i n terms of the total amount of c arbon dioxide (CO 2 ) emission s, and ranked 2nd (around 145 percent ) among them in growth of emissions, from 1990 to 2014 (OECD, 2017 ). 61 Similarly, regarding the amount of CO 2 emission per capita and its growth rate from 1990 to 2014, Korea ranked 6th and 1st among thirty five OECD member coun tries, respectively (OECD, 2017 ). Overall, while South Korea has enj oyed economic growth, it now faces challenges in alleviating GHG emissions. Literatures related to Climate Change Risk Perception in South Korea A cc ording to t he 2013 national survey on environmental conservation conducted by the Ministry of Environment (2014 ), in South Korea, the general public indicate s climate change as the second m ost serious problem among ten environmental issues (e.g., household waste, toxic materials, and water poll ution) , following industrial waste. Also , regarding the improvement 60 Data are expressed constant 2010 U.S. dollars. 61 Worrell et al ., 2001, p. 304). According to OECD (2017), greenhouse gases include seven gases, such as carbon dioxide (CO 2 ), methane (CH 4 ), nitrous oxide (N 2 O), chlorofluorocarbons (CFCs), and hydrofluorocarbons (HFCs).

PAGE 80

66 priority of environmental issues , climate change is ranked third among the ten proble ms , fol lowing the natural environment and water. 62 While many Koreans perceive climate change as a serious and urgent environmental problem , there are very few studies that deal with climate change ( risk ) perception in South Korea. Also, these existing studies mostly focus on particular fields , such as education ( Seo et al ., 2013) , agriculture and fisheries ( Kim, Lee, & Jeong, 2014; Park, Lee, & Kim, 2014) , and tourism ( Ahn, Lee, & Hwang, 2016; Kim & Kim, 2013; Lee, Hwang, & Seo, 2015; Min, Kim, & Kim, 2015), or certain sub national locations (Choi, Park, & Jeon, 2014; Kim, Kim, & Lee, 2014 ; Koh & Yi, 2016). 63 In other words, there is no clear evidence about the causal relationship among persona l factors, risk perception, and behavioral responses related to climate change at the nation al level. Thus, t his study will become a stepping stone for future research investigating climate change risk perception and behavioral responses of the general public in South Korea . Also, t he findings of this national level analysis could help policymakers design effective climate policies in both South Korea and countries with similar socio economic conditions . Research Methods Data Collection Research tool . Data were collected from a web based survey implemented in South Korea. Climate change risk perception and climate policy assessment formed by a psychological process are typically invisible. Personal experiences, knowledge, and behavioral r esponses to 62 Since 1982, the Ministry of Environment has conducted surveys associated with the national consciousness on environmental conservation in South Korea, typically once every five years (e.g., 2003, 2008, and 2013). The respo ndents of the 2013 survey consist of 1,500 citizens and 200 environmental experts. 63 Kim, Kim, and Lee (2014) investigated public awareness of climate change in Chungnam province, located in the west of South Korea . Choi, Park, and Jeon (2014) studied publ ic perception of climate change in Gangnam gu,

PAGE 81

67 mitigate clim ate change at the national level cannot be easily measured by experiments and field studies. The majority of empirical research projects intr oduced in the literature review chapter s depend on survey (e.g., questionnaires or interviews) data or secondary survey data. Therefore, this study obtains these personal and ha rd to observe da ta at the national level from the Korea Water Resources Corporation (K water) K water conducted the survey for establishing South . While Web based surveys can reduce cost s , save time, and offer flexibility in the questionnaire design, this method has a major weakness, called coverage error Singleton & Straits, 2017, p. 271). 64 As of 2016, the proportion of total fixed and wireless bro adband subscriptions in South Korea was over 100% (Natio nal Information Society Agency , 2016; OECD , 2016) , and about 90% of the entire population utilize smart phones (Statistics Korea, 2017a). Also, large scale Web based survey platforms, such as ls) , operat e in South Korea (Open Survey , 2017; Marketlink, 2017). Overall, South Korea has a favorable environment for a Web based survey with less coverage error. Survey design and refinement . Good survey instrument design is one of the key ingredients to the success of survey research (Singleton & Straits, 2017 ). 65 A questionnaire was developed by researchers at the research center for water p olicy and e conomy in the K water institute , who have substantial experience conducting climate (or water) policy related studies . 64 based, Web enabled panels designed to be representative of target po ( Singleton & Straits, 2017, p. 280). 65

PAGE 82

68 Singleton and Straits ( 2017) , in evaluating drafts of survey questions, highlight that p. 336 ) is a prerequisite for minimizing the risk of study failure. To evaluate and improve the design of the questions, expert reviews of the survey questionnaire were carried out. Based on the results of the expert reviews, the questionnaire was refined. Key question s in the questionnaire reflect the following topics: a) climate change risk perception (e.g., the seriousness of the adverse impacts of climate change ); b) experiences of extreme weather events (e.g., experiences of property damage by drought and flood); c) experiences of climate change (e.g., the rising of average temperature, the change in precipitation, and the increase of the danger of flood and drought); and d) behavioral responses to mitigate climate changes (e.g., saving energy, reducing the use of disposable items, and utilizing renewable energy). Appendix B presents the key questions of the questionnaire. Most questions are closed ended , mainly in the form of a four point scale. 66 Representativeness . This self administered survey was carried out by the online survey on August 10, 2016 and had 1,500 respondents who were randomly selected among its panels. The target population was individuals who reside in South Korea, 51.7 million as of 2016 (Statistics Korea , 2017b) . These respondents represent a good cross section of gender , age, and region in South Korea. 67 T he socio demographic characteristics of the respondents compared to those of the South Korean population are shown in Table 2 . 66 According to Singleton and Straits ( 2017), while it is difficult to develop closed ended questions, they require less respondent effort and facility with words than open respondents with respect to s ome well ). 67 This survey appears to oversample more educated people compared to the population. However, the proportio n of people with higher education (i.e., undergraduate and above) among the entire population is increasing quickly (e.g., from 40% in 2011 to 47% in 2016) (Ministry of Education, 2018).

PAGE 83

69 Table 2 . 2016 s urvey r socio demographic c haracteristics Classification Respondents South Korean Population Num. Per. Num. Per. Gender Male 755 50.3% 25,827,594 50.0% Female 745 49.7% 25,868,622 50.0% Age (0 19) (10,047,206 ) 20 29 257 17.1% 6,758,428 16.2% 30 39 292 19.4% 7,534,140 18.1 % 40 49 343 22.9% 8,797,053 21 .1 % 50 59 325 21.7% 8,424661 20.3 % 60 283 18.9% 10,134,728 24.3 % Education Under middle school 32 2.1% 13 .0 High school 416 27.8% 40.0 U ndergraduate 888 59.2% 47 .0 Graduate and above 164 10.9% Income (Average Monthly Income before Tax) I
PAGE 84

70 Variable Description Mediating and dependent variable s . Climate change risk perception and behavioral responses to address climate change are employed as a mediating (both dependent and explanatory) variable and a dependent variable , respectively. Climate change risk perception . This study utilizes a single item, the seriousness of climate change, as an indicator of climate change risk perception . According to PMT , threat appraisal is conducted based on the two following questions: a ) does an individual believe that the threat is severe? (i.e., perceived severity) ; and b) does an individual feel vulnerable? (i.e., perceived vulnerability). This study considers that perceived seriousness (i.e., the quality of worrying because of possible risks) of climate change can embrace both perceived severity and vulnerability of climate change. For example, if an individual perceives the higher level of severity and the lower le vel of vulnerability related to climate change, the level of his or her perceived seriousness of climate change may be moderate, and vice versa. If an individual perceives the higher levels of both severity and vulnerability associated with climate change, the level of his or her perceived seriousness of climate change also may be higher, and vice versa. Respondents were asked to select their level of seriousness on the adverse impacts of climate change. They presented their level of seriousness for the que stion on a four point scale, where 1 was strongly serious and 4 was not serious . This study reverses the scale of the question . Thus , higher values indicate greater perception of the seriousness of the problem. Behavioral responses to address climate change . This paper uses the number of behavioral responses to mitigate climate change as a dependent variable. Considering the definition of behavioral responses to climate change in Appendix A, t his variable include s not only direct behavioral efforts to reduce GHG emissions (e.g., saving energy and utilizing

PAGE 85

71 renewable energy ), but also indirect behavioral efforts (e.g., donations to environmental organizations) to mitigate climate change. Based on a single question, r espondents were asked to indicate all activities they were practicing to address climate change among the following eight activities : a) saving energy; b) saving water; c) using public transportation instead of a privately owned car; d) reducing the use of disposable items; e) utilizing renewable energy; f) enjoying a vegetarian diet instead of a meat diet; g) donations to environmenta l organizations; and h) voting based on environmental related election promises. Recycling, which has usually been utilized by previous studies (e.g., Masud et al ., 2016; Semenza et al ., 2008; Whitmarsh , 20 09a), is excluded from the activities because it is required by law in South Korea ( Ministry of Government Legislation, 2015 a ). The number of practicing activities are coded on a nine point scale ( from 0 to 8). 68 Additionally , the practice of each activity is employed as a sub dependent variable to reveal more details of the relationship between risk perception and behavioral responses . The answers of the respondents are divide d by each activity , and the practice of each activity is dichotomously coded , where 0 = no and 1 = yes . Independent variables . T hree experience variables , three knowledge variables , and two policy as sessment variabl es are utilized as ind ependent variable s . Short term experiences . Given that drought and flood are most commonly regarded as significant weather events intensified by climate change, the experiences of drought and flood are employed as the indicators of short term experiences . 69 To clarify the range of drought and 68 While each activity has the potential to mitigate climate change, i t is very challenging to compare the impacts of those activities. Therefore, this study posits that they have similar effects on climate change mitigation (i.e., giving the same weight to each response), and it might not be highly problematic. 69 The heat w ave is also viewed as one of the significant weather events associated with climate change, ( Akompab et al ., 2013; Beniston & Diaz, 2004; Hansen, Sato, & Ruedy, 2012) . However, it is not dealt with in this analysis due to a lack of data.

PAGE 86

72 flood experiences, this study focuses on experiences of property damage by drought and flood . Respondents were asked to choose their answers among four items, where 1 = no experience , 2 = by drought , 3 = by flood , and 4 = by both drought and flood . E xpe riences of property damage by drought (hereafter the experience of drought) is measured as a dichotomous variable, where no = 0 and yes = 1 (those who select ed 2 or 4). Also, expe riences of property damage by flood (hereafter the experience of flood) is co ded in the same way, where no = 0 and yes = 1 (those who chose 3 or 4). Long term experiences . Climate change has long term features such as variations in temperature and precipitation . To measure the experiences of those features , r espondents were asked to indicate their level of experiences on the rising of average temperature, the change in precipitation, the increase of the danger of flood, and the increase of the danger of drought in the last ten years, respectively. They point ed out their level of experiences for each question o n a four point scale, where 1 = definitely , 2 = somewhat , 3 = little , and 4 = not . Th e scale s of the question s are reverse code d , where 0 = not , 1 = little , 2 = somewhat , and 3 = definitely . T hose values are summ ed into a single measure, the index of climate change experiences alpha of this index is 0.6982 . Also, t his study tests whether those four items all fall along one dimension. According to principal component factor analysis ( the critical point of Eigenvalue is 1) , the number of retained factor is one. Specifically, the Eigenvalue of the first factor is 2.16. This is fifty four percent of t he total possible of 4 . By comparison, that of the sec ond factor is just 0.74. All correlatio n values are greater than 0.3 except correlation between the rising of average temperature and the increase of danger of flood (0.27). The overall value of Kaiser Meyer Olkin (KMO) measure is 0.7384 , and it is located in the middling

PAGE 87

73 (Kaiser, 1974) . 70 Lastly, the result of Bar significant at the p<.001 level. Overall , those four items are adequate for creating the index of climate change experiences . The range of this index is from 0 to 12 . In other words , the higher the index is, the long term climate change is . Knowledge . Two general knowledge variables (i.e ., awareness of climate change and the understanding of the cause of climate change) and one specific knowledge variable (i.e., the familiarity with a go vernment GHG reduction plan) were measured. 71 First, awareness of climate change is simply coded as a dichotomous variable where no = 0 and yes = 1. Second, regarding the understanding of the cause of climate change , respondents were asked to present the contribution of human activities to climate change by numeric scores from 0 up to 100. The former mean s that climate change has nothing to do with human activities (e.g., G HG emissions ), and the latter mean s that climate change is caused mainly by human activities. Given the scientific evidence identifying GHG emissions as the main cause of climate change , t hose who indicate higher scores have a higher understanding of the cause of climate change than those who indicate lower scores. Lastly, the reduction of GHG emissions is a core strategy of governments to mitigate climate change. R level of familiarity with a government GHG reduction plan was measured o n a four point scale, where 1 was precisely and 4 was not . This study reverses the scale of the question: Higher values indicate greater familiarity with the government GHG reduction plan. 70 0.00 to 0.49; miserable 0.50 to 0.59; mediocre 0.60 to 0.69; middling 0.70 to 0.79; meritorious 0.80 to 0.89; and marvelous 0.90 to 1.00. 71 The understanding of th e consequences of climate change is also one of the important components in climate change knowledge (see Appendix A). Unfortunately, there is no proper item for measuring that component in the survey data.

PAGE 88

74 P olicy assessment . As alluded to above, drought and flood are extreme weather events deeply associated with climate change . Therefore, policies to address these events are important part s of government s climate change. assessment of drought policy a nd flood policy, respondents were asked to answer their level of evaluation for each policy on a five point scale, where 1 = very poor , 2 = poor , 3 = fair , 4 = good , and 5 = excellent . 72 Control variables . Five socio demographic v ariables are included to isolate the influence of personal factors on climate change risk perception and behavioral responses to mitigate climate change. First, gender is coded as a dichotomous variable, where 0 = male and 1 = female . Second, this study includes age expressed in years . Third, the level of education is measured on a four point scale, where 1 = under middle school , 2 = high school , 3 = undergraduate , and 4 = graduate and above . Fourth , this analysis includes the level of income (i.e., average monthly income before t ax last year) as a seven point scale variable, range from 1 for those earning less than KRW 1million ( around US$ 909) to 7 for those earning over KRW 6million ( around US$ 5,455) . Lastly, residential area is coded on a three point scale, where 0 = me tropolitan government or cities , 1 = middle or small cities , and 2 = counties . 73 Data Analysis Given that this study focuses on the relationship among personal factors, risk perception, 72 Considering the conceptual map in Appendix A, t hese policies should be specifically evaluated on their necessities, validities, efficiencies, etc. However, in reality, it is difficult for the general public to precisely assess t he policies on those points. Also, the survey includes a single item for ea ch policy assessment . This study assumes that the overall evaluations for each policy can represent an indicator of policy assessment and will not be problematic. 73 In South Korea, the populations of counties, cities, and metropolitan cities are generally less than 100,000, more than 100,000 and less than 1,000,000, and more than 1,000,000, respectively (Ministry of Government Legislation, 2017b; Statistics Korea, 20 19).

PAGE 89

75 (Montgomery, Peck, & Vining, 2012, p. 1) , is a suitable tool to evaluate data. Among diverse regression methods (e.g., linear, logistic, and structural), this study employ s Structural Equation Modeling (SEM) ( Goldberger, 1972) . While most regression methods allow analysis of the relationship between only one layer of linkages between independent and dependent variables at Gefen, Straub, & Boudreau, 2000 p. 4), SEM , including confirmatory factor analysis, path analysis, and others , enables to analyze the relationship among all the variables in the model simultaneously (Gefen, Straub, & Boudreau, 2000; Kline, 2015; Schumacker & Lomax, 2004) and to produce better performance in systematic and comprehensive analyses rather t han other regression models ( Iacobucci, Saldanha, & Deng, 2007) . By utilizing SEM, not only can this study conduct simultaneous analysis of the dynamic relationship among the variables, but also identify the direct and indirect effects of the personal fact ors on climate change risk perception and behavioral respons es to mitigate climate change. This ana l ysis is composed of four phases. First, socio demographic characteristics and descriptive results are introduced. Second, biv ariate correlations between dependent and other variables are computed. Third, prior to SEM analysis, model fit is evaluated based on the values of chi Approximation (RMSE A), Standardized Root Mean Squared Residual (SRMR), Comparative Fit Index (CFI), and Tucker Lewis Index (TLI). 74 Lastly, SEM regression analysis is operated by 74 According to Hu and Bentler (1999), the primary goals in the application of SEM are the assessment of methods to assess model fit: The chi Mullen, 2008; Hu & Bentler, 1999) categorize fit indices into two types: absolute fit indices and incremental fit indices, also known as relative fit indices ( McDonald & Ho, 2002). Hu and Bentler, 1999, p. 2), includes the Goodness of Fit Index (GFI; Jöreskog & Sörbom, 1984), Root Mean Square Error of Approximation (RMSEA; Steiger & Lind, 1980), Standardized Root Mean Squared Residual (SRMR; Bentler, 1995), etc. The Chi Square ( also included in this type. The

PAGE 90

76 using S SEM package and an appropriate estimation method . Pathways in the national level research model are presented in Figure 5. In addition to the main model, eight s ub models that utilize each behavioral response coded dichotomously as a dependent variable are analyzed by Generalized Structural Equation Modeling (GSEM) , and the results are reported in the alternative model specifications section. 75 Figure 5 . Pathways in the national level research model Before finalizing the research model, several pilot models, including pathways amon g personal factor variables, were tested. These pilot models included causal relationships among the subset of personal factor variables, such as between short term experience and policy assessment (e.g ., more drought experiences could be negatively associated with the level of Hu and Bentler, 1999, p. 2), consists of the Comparative Fit Index (CFI; Bentler, 199 0), Normed Fit Index (NFI; Bentler & Bonett, 1980), Tucker Lewis Index (TLI: Tucker & Lewis, 1973) , etc. 75 Specifically, these sub models are analyzed by GSEM logit. GSEM regression analysis in STATA does not provide model fits.

PAGE 91

77 drought policy asse ssment) and between general knowledge and policy related knowledge (e.g., greater general knowledge of climate change could be positively associated with the level of policy relat ed knowledge) . However, while these models may help understand the relationship among the personal factors, their model fits are very poor and not acceptable. 76 Th e final models therefore exclude pathways that examine the inter relationship s among the personal factor variables in unveiling related to climate change. Analysis Results Socio demographic Characteristics The responden ts are from all regions (i.e., seven metropolitan cities, eight provinces, and two self governing areas) in South Korea. They range in age from 20 to 87 years (M = 44.40, SD = 13.62) and 49.7% are female s advanced degree. Approximately 65.5% earn over KRW 3million (US$ 2,727) and a bit more than a quarter earn over KR W 5million (US$ 4,545) a month. Descriptiv e Results Most of the respondents are aware of climate change (M = 0.96, SD = 0.21) and view human activities as the key driver of climate change (M = 75.37, SD = 16.36) . When asked to consider the impacts of climate change, nearly 98% of the respondents ans wer ed that the negative impacts of climate change (M = 3.55, SD = 0.55) are either strongly serious (57.1%) or serious (40.9%) . Regarding personal experiences, 11.6% and 21.9% of the re spondents have 76 For example, the pilot mo del adding the pathways between short term experiences and policy assessment and between general knowledge and policy = 35, p<.001), RMSEA 0.150, CFI 0.269, TLI 0.253, and SRMR 0.079). Also, while drought experience and awareness of climate change significantly and positively affect drought policy assessment and familiarity with a government GHG reduction plan respectively, flood experience and understanding of the cause of clima te change do not have a significant relationship with flood policy assessment and familiarity with a government GHG reduction plan, respectively.

PAGE 92

78 experiences of drought and flood, respectively. Meanwhile, the majo rity of the respondents answer that they ( either definitely or somewhat ) experience climate changes in the last ten years. As shown in Figure 6 , t he risin g of average temperatures (M = 2 .80, SD = 0.45), th e change in precipitation (M = 2 .52, SD = 0.62), and the increase of the dange r of drought (M = 2 .55, SD = 0.61) a re experienced by more than 90%, respectively. The increa se of the danger of flood (M = 2 .08, SD = 0.80) is experienced by 75.7 %. Figure 6 . R esults of four long term climate change experiences Concerning the while 38.1 % and 47.7% of the respondents negatively (either poor or very poor ) evaluate the flood and drought p olicies , respectively , o nly 8.5% and 6.3% of the respondent s positively (either good or excellent ) assess them, respectively. Many of the respondents are on neutral ground ( fair ), 53.4% for the flood policy and 46.0% for the drought policy. The proportion of the respondents who are familiar with the g over nment GHG reduction plan ( either moderately or precisely ) (44.8%) is slightly less than that of the respondents who are not ( either har dly or not ) (55.2%) ( M = 2.33, SD = 0.84 ). Lastly, as shown in Figure 7 , the respondents are practicing around three behavioral responses to address clim ate change on average ( M = 3.49, SD = 1.48).

PAGE 93

79 Figure 7 . R Bivariate Cor relation Coefficients among Variables Table 3 presents t he inter correlations of the key variables in the research mode l. In the case o f the relationship between climate change risk perception a nd independent variables (Table 3 .1) , all correlations between them show the hypothes ized directions (i.e., three experience and two knowledge factors positive; two policy assessment factors negative). However, climate change risk perception does not have a statistically significant correlation with short term experiences , whereas other re lationships are significant at the p<.001 level. Amo ng socio demographic variables, gender and age have a significant positive relationship with climate change risk perception at the p<.001 and p<.05 levels, respectively. In terms of the correlations between behavioral responses to mitigate climate change and other variables , including climate change risk perception (Table 3 .2) , the number of behavioral response s are significantly correlated with climate change risk perception and three personal variables at the p<.001 level , and all these correlations show positive directions. Among socio demographic variables, age and income have a significant positive relationship with the behavioral response s at the p<.001 and p<.05 levels, respectively.

PAGE 94

80

PAGE 95

81 M odel Fit While the Chi Square value ( has traditionally been measured for evaluating model fit for SEM , it has a critical limitation, sensitivity to sample size (i.e., mostly rejecting the model with large samples) ( Hox & Bechger, 2007; Hu & Bentler, 1999). Therefore, diverse fit indices, such as Root Mean Square Error of Approximation (RMSEA), Standardized Root Mean Squared Residual (SRMR), Comparative Fit Index (CFI), and Tucker Lewis Index (TLI), have been proposed by researchers to supple ment the Chi Square ( 2008; Hox & Bechger, 2007; McDonald & Ho, 2002). Regarding the research model, this study p< .01), RMSEA = .038 (90% CI), SRMR = .010, CFI = .970, and TLI = .867. Given acceptable thresholds of fit indices in the previous studies, the research model fits the data well. 77 R egression Analysis This study hypothesizes that personal factors influence cl imate change risk perception ( H1 , H2 , and H4 ), and some personal factors and climate change risk perception lead to behavioral responses to address climate change ( H1 , H3 , and H5 ) . SEM is utilized to test the hypotheses about the relationship among personal factor, risk perception, and behavioral responses related to climate change . Prior to analysis, the selection of an appropriate estimation method is needed. Most SEM related programs set Maximum likelihood (ML) as a default method due to the precision of estimates (smallest variance) ( Ullman & Bentler , 2003 ). 77 Based on previous studies, Hooper, Coughlan, and Mullen (2008) [1] and Hu and Bentler (1999 ) [2] provide the acceptable thresholds (conventional cutoff values) of fit indices. Those are as follows: Fit indices Acceptable thresholds (cutoff values) Chi [1] An insignificant p value (p > .05) Root Mean Square Error of Approximation (RMSEA) [ 1] Values less than .07 / [2] less than .06 Standardized Root Mean Squared Residual (SRMR) [1] [2] Values less than .08 Comparative Fit Index (CFI) [1] [2] Values greater than .95 Tucker Lewis Index (TLI) [1] [2] Values greater than .95

PAGE 96

82 However, an important assumption for ML estimation is the full joint normality of all the variables (i.e., multivariate normality) ( Chou & Bentler, 1995; Finney & DiStefano , 2006 ; Ullman & Bentler , 2003; West, Finch, & Curran, 1995 ) . 78 U nivariate and multivariate normality tests are conducted for variables in the research model. While all univariate skewness and kurtosis of each variable except one variable (i.e., awareness of climate change) are placed beyond the range of non normality concern (i.e., |skewness| > 2, |Kurtosis| > 7) suggested by previous studies ( multivariate ( skewness and kurtosis) test indicat e that multivariate normality assumption is not satisfied. Namely , the normality assumption is violated. Considering several approaches for non normality (see F o otnote 78 ) , the results of SEM regression analyses with different estimation methods (or options) (i.e., ML, ML with robust standard errors, ML with bootstrapping, ADF, and Generalized SEM) are presented in Appendix C. This appendix includes the results of GSEM, specifically Poisson regression which fits models with count outcome variables ( Coxe, West, & Aiken, 2009) , by assuming that the dependent variable (i.e., the number of behavioral responses) is a count variable. Overall, directions and significance of coefficients are almost identical. Therefore, this study tries to answer the research question by interpreting the results of the SEM regression analysi s using ML estimation with robust standard errors. 79 The results of 78 When violating the normality assumption, there are several approaches to dealing with non normality, such as ML with the Satorra Bentler scaled chi squares and robust standard errors ( Satorra & Bentler, 1988; 1994), bootstrapping (Efron, 1979; Yung & Bentl er, 1996 ; Finney & DiStefano , 2006 ), and Asymptotically Distribution Free (ADF) (Browne, 1984) . While ADF makes no assumption of normality, it does not show good performance with sample size less than 2,500 ( Hu, Bentler, & Kano, 1992; Ullman & Bentler , 2003). Bootstrapping, which refers to is viewed as a promising method to correct standard errors (Nevitt & Hancock, 2001, p.355 ) . However, it may also present erratic results with small sample size (e.g., 200 or less) (Nevitt & Hancock, 2001). Generally, ML with the Satorra Bentler scaled chi squares and robust standard errors is regarded as a good approach to handling non normality (Curran, West, & Finch, 1996; Hu, Bentler, & Kano, 1992). 79 There is no evidence of multicollinearity, heteroscedasticity, influential outliers, and other significant statistical problems.

PAGE 97

83 the SEM regression analysis and significant pathw ays in the research model are described in Table 4 and Figure 8 , respectively. Also, direct, indirect, and total effects of independent variables on behavioral responses to alleviate climate change are presented in Appendix D . Table 4 . R esults of SEM regression analysi s about the relationship among personal factor s , risk perception, and behavioral resp onses related to climate change Dependent Variables Independent Variables Coeff. Robust Std. Err. Climate change risk perception Drought experience .049 .039 Flood experience .001 .032 Climate change experience .108*** .008 Awareness of climate change .216** .079 Understanding of the c ause of climate change .005*** .001 Assessment of drought policy .065** .021 Assessment of flood policy .021 .022 Gender (female) .132*** .026 Age .002 .001 Education .008 .022 Income .005 .008 Area .003 .021 Constant 1.830 The number of b ehavioral responses Climate change risk perception .260*** .069 Drought experience .300 .156 Flood experience .118 .099 Familiarity with a government GHG reduction plan .206*** .049 Gender (female) .078 .075 Age .017*** .003 Education .027 .060 Income .044 .024 Area .036 .063 Constant 1.147 n 1,484 Note : Sixteen respondents who did not answer at least one question out of the four questions related to the experiences of climate change are excluded from the SEM regr ession analysi s . *p<.05, ** p<.01, ***p<.001 (two tailed test)

PAGE 98

84 *p<.05, ** p<.01, ***p< .001 (two tailed test) Figure 8 . Significant pathways in the national level research model Climate change risk perception. Four independent and one socio demographic variables show a statistically significant relationship with climate change risk perception. Consequently, the results fully identify Hypothesis 2 and partially verify Hypothesis 1 and 4 . E xpe rience . Long term climate ch ange experiences have a significant positive relationship with climate change risk perception (b = . 108, p<.001). This result indicates that t he respondents with more intense climate change experiences are likely to have higher levels of climate change risk perception and corroborates the findings of Akerlof et al . (2013), Donner and McDaniels (20 13), and Li, Johnson, and Zaval (2011) . Contrary to expectation, t wo short term experiences do not significantly influence climate change risk perception . G eneral knowledge . A s hypothesized , both awareness of climate change (b = .216) and the understanding of the cause of climate change (b = .005) have a positive correlation with cli mate change risk perception at the p<.01 and p<.001 level s, respectively . T he more the

PAGE 99

85 respondents are aware of climate change or understand the cause of climate change , the higher level s of climate change ris k perception they have. These results are not surprising as previous studies (e.g., Lee et al ., 2015; Milfont, 2012; Shi et al . , 2016) report that more knowledge about climate change, such as the cause a nd consequence of that , is correlated with greater concern about climate change. Policy assessment . While this study predicts that the ass essments of both drought and flood policies would negatively influence climate change risk perception, only t he assessment of drought policy negatively affect s climate change risk perception . T he lower the higher levels of climate ch ange risk perception they have. This result provides partial evidence to support the argument of Grothmann and Patt (2005) that increase if climate policies implemented by governments fail to alleviate climate change risk. Socio demographic variables . Only gender (b = .132) has a positive relationship with climate change risk per ception at the p<.001 level . Among the respondents, women have higher levels of climate change risk perception than men. This finding supports other research projects which find that women tend to be more concerned about environmental problems ( Blocker & Eckberg, 1997; . 80 Behavioral responses to mitigate climate change . Five independent and two socio demographic variables directly or indirectly (i.e., via climate change risk perception) affect to mitigate climate change . Based on the results of the SEM regression analysi s, Hypothesis 3 and 5 are fully substantiated . 80 However, even though numerous research projects have investigated gender difference in environmental concern, this topic is still controversial. the upshot is a mix of both robust patterns and 66).

PAGE 100

86 Climate change risk perception . T his analysis verifies the significant and positive relationship between climate change risk perception and behavioral responses t o mitigate climate change (b = .2 60, p<.001) . The respondents with greater climate risk perception are more likely to take behavioral responses to alleviate climate change. This finding is consis tent with many existing studies , which show climate change risk perception to be a significant motivator of behavioral responses to cope with climate change (e.g., Deng, Wang, & Yousefpour, 2017; Spence et al . , 2011 ; Yu et al ., 2013) . Long term climate change experience s, two knowledge variables (i.e., awareness of climate chan ge and the understanding of the cause of climate change), and the assessment of drought policy significantly and indirectly affect behavioral responses to mitigate climate change via climate change r isk perception (see Appendix D) . S h ort term experience s . Unexpected ly, t here are no significant direct effect s of short term experience s on behavioral responses to mitigate climate change . As a result , b oth d rought and flood experience s fail to influence behavioral response s through direct and indirect path s . 81 Overall , these findings are not in line with the empirical evidence o r arguments of previous studies that show direct experiences of extreme weath er events have a positive impact on risk perception or precautionary (or preparedness) actions ( Demski et al ., 2017; Deng, Wang, & Yousefpour, 2017; et al ., 2005; Zaalberg et al ., 2009 ) . However, t his lack of impact from direct ex periences is also shown in some other studies (e.g., Carlton et al ., 2016; Martin, Martin, & Kent, 2009). This dissertation considers that t he relationship between 81 The pilot model, which included three types of short term experiences (i.e., only drought, only flood, and both drought and flood), showed similar results. No significant relationship between short term experiences and climate change risk perception and between short term experiences and behavioral responses to mitigate climate change was found except the weak relationship between both drought and flood experience and behavioral responses to mitigate climate change (b = .434. p<.05).

PAGE 101

87 sho r t term experience s and behavioral responses to mitigate climate change sho uld be more deeply investigated, and it will return to this relationship in the expert level analysis. Policy related knowledge . T he familiarity with a g overnment GHG reduction plan shows a significant positive correlation with behavioral responses (b = .206, p<.001 ) . T h is novel finding identifies that the respo ndents with more climate policy related knowledge have a tendency to take more behavioral responses to curb climate change . Socio demographic variables . A ge shows a positive relationship with behavioral responses through a direct path ( b = .017, p<.001) . T he elderly are more likely to take behavioral responses to mitigate climate change than the youth. This finding supports several studies ( Keshavarz & Karami, 2016; Rainear & Christensen, 2017) and contradicts the argument of Semenza et al . (2008) that young people are more likely to change their behaviors to mitigate climate change than elderly people . With regard to other variables, only gender indirect ly influences behavioral responses (b = .034, p<.01) . Among the respondents, women tend to be more engage d in behavioral responses to mitigate climate change than men . 82 E du cation , income, and area do not significantly affect behavioral responses both directly and indirectly. Alternative Model Specifications Regression analyses of sub models. E ight sub models, employing each dichotomous behavioral response (0 = no and 1 = yes ) as a dependent variable, are investigated by GSEM regression analysis . The results of the analyse s are presented in Table 5. The findings complement those derived from the main model by unveiling the several noteworthy patterns of the relationship between behavioral responses and core independent variables. 82 Mean while, Blocker and Eckberg (1997) take a different stance. They mention that while women generally have greater concern about the environment than men, it is unlikely that they take more pro environmental actions than men.

PAGE 102

88 First, climate change risk perception has a much greater impact on direct behavioral responses to mit igate climate change, such as saving energy, reducing the use of disposable items, and utilizing renewable energy , rather than indirect behavioral responses such as saving water, donations to environmental organizations, and voting based on environment al r elated election promises . U sing public transportation instead of a privately owned car is the sole direc t behavioral response not affected by climate change risk perception, and it is negatively influenced by area. The respondents who live in urban areas are more likely to use public transportation instead of a privately owned car than those who live in rural areas . Given higher levels of public transportation infrastructure and traffic congestion in urban areas, this result is understandable , and their im pact s are likely to make inroads on the influence of c limate change risk perception to this response. Second , the familiarity with a government GHG reduction plan intensifies three indirect behavioral responses , but only one direct behavioral response. Considering the necessity of both direct and indirect behavioral responses in addressing climate change, these result s imply that policy related knowledge and risk perception are helpful to one another. Third, the relationship between drought experience an d each response shows contradictory results. While four activities, such as utilizing renewable energy and donations to environmental organizations, are positively influenced by drought experience, two activities, such as saving energy and saving water, ar e negatively affected. The se conflicting impacts might lead to the insignificant relationship between drought experience and the number of behavioral responses in the main research model . T his dissertation does not try to reveal more details of this phenomenon due to a lack of data . L ast , age has a significant impact on both direct and indirect behav ioral responses. Five responses are intensified by age. Also, gender, income , and area are related to a couple of activities. Overall, the results in the sub models enrich thos e in the research

PAGE 103

89 model, and t he identified patterns can help both practitioners and researchers pra ctically and academically . Table 5 . R esults of GSEM regression analysis about the relationship among personal factors, risk perception, and eac h behavioral response associated with climate change Independent variables Dependent variables (each behavioral response) Saving energy Using public transport Reducing the use of disposable items Utilizing renewable energy Saving water Enjoying a vegetarian diet Donations to ENV. ORG. Voting based on election promises Direct behavioral responses for mitigation In d irect behavioral responses for mitigation Coeff. Coeff. Coeff. Coeff. Coeff. Coeff. Coeff. Coeff. Climate change risk perception .323** (.122) ( 1.382**) .025 (.099) (.976) .299** (.111) (1.348**) .408** (.126) (1.504**) .196 (.117) (1.196) .283* (.133) (1.327*) .021 (.175) (.980) .217 (.142) (1.242) Drought experience .473* (.213) (.623*) .044 (.179) (.957) .373 (.199) (.689) .848*** (.194) (2.335***) .407* (.198) (.666*) .593** (.207) (1.809**) 1.008*** (.235) (2.739***) .721*** (.204) (2.057***) Flood experience .249 (.173) (.780) .173 (.136) (1.188) .012 (.156) (1.012) .157 (.166) (1.170) .017 (.163) (.983) .146 (.166) (1.157) .243 (.219) (1.275) .313 (.172) (1.368) Familiarity with a GHG reduction plan .122 (.088) (1.129) .025 (.067) (1.025) .149 (.077) (1.161) .343*** (.087) (1.409***) .022 (.080) (.979) .232* (.092) (1.261*) .442*** (.119) (1.555***) .329*** (.094) (1.390***) Gender .168 (.140) (.845) .269* (.109) (1.308*) .449*** (.127) (1.567***) .491*** (.141) (.612***) .002 (.129) (1.002) .268 (.139) (1.307) .071 (.182) (1.074) .172 (.148) (.842) Age .027*** (.005) (1.028 *** ) .002 (.004) (.998) .021*** (.005) (1.021***) .014** (.005) (1.014**) .020*** (.005) (1.021***) .033*** (.005) (1.034***) .002 (.007) (1.002) .008 (.006) (.992) Education .069 (.117) (1.072) .070 (.086) (.932) .133 (.098) (.875) .146 (.109) (.864) .037 (.103) (.964) .090 (.109) (1.095) .112 (.157) (1.119) .048 (.120) (1.049) Income .057 (.041) (1.059) .007 (.033) (.993) .134** (.039) (1.144**) .034 (.043) (.966) .060 (.040) (1.061) .016 (.044) (1.016) .025 (.058) (1.026) .015 (.046) (1.015) Area .027 (.109) (1.027) .301*** (.086) (.740***) .078 (.099) (.925) .083 (.111) (.920) .042 (.104) (1.043) .146 (.114) (1.157) .013 (.143) (1.013) .241* (.121) (1.273*) Constant 1.440 .553 1.595 3.655 3.655 5.218 4.023 3.386 n 1 , 484 1 , 484 1 , 484 1 , 484 1 , 484 1 , 484 1 , 484 1 , 484 Note 1. The relationship between personal factors and risk perception is omitted to avoid duplication. 2. The first and second parentheses display robust standard errors and odds ratios, respectively. 3. Multicollinearity, influential outliers, specification errors, and other statistical problems are not detected. *p<.05, ** p<.01, ***p<.001 (two tailed test)

PAGE 104

90 Relationship between s ocio demographic characteristics and personal factors . When looking at the research model , personal factor variables are described as exogenous variables. However, as mentioned earlier, there could be a causal relationship among the per sonal factor variables (see pp. 76 77 ). Also, even though five socio demographic characteristics are only utilized as the control variables of risk perception and behavioral response variables in the research model, some may have a causal relationship with these personal factor variables. Table 3 already showed the several significant correlation s between the personal factor and socio demographic characteristic variables. Intuitively, the level of educatio n and age may positively affect knowledge and experience variables, respectively. It is not surprising that t he results of regression analyses about the relationship between each personal factor variable and five socio demographic characteristic variables in Appendix E are mostly consistent with those expectations . Nevertheless, this dissertation leaves these rel ationships out of the research model because of the following reasons. First, identifying the causal relationship between them may not be a significant and innovative contribution to the development of climate change literature. Second, the impacts of the socio demographic characteristic variables, mediated by the personal factor and risk perception variables successively, on the behavioral response variable would be minimal. Third, it is difficult and ineffective for policymakers to implement policies that can demographic characteristics to promote th eir behavioral responses against climate change. Last , when the research model includes these relationships, its model fits are greatly undermined and become unacceptable . Overall, revealing the causal relationship between the personal factor s and socio demographic characteristic s is somewhat meaningful, but not critical to this study .

PAGE 105

91 Reverse pathway between risk perception and beh a vi o ral responses . Contrary to this study core hypothesis that climate change risk perception positively affects behavioral responses to address climate change, some studies (e.g., Bem, 1965; Kiesler, 1971; Olson & behaviors. To ensure the appropriateness of the research model, the reverse relationship between risk perception and behavioral responses related to climate change is carefully checked by employing the following two approaches: a) switching the boxes of the risk perception and behavioral response variables (i.e., personal factors in Figure 5, leaving the pathways uncha nged ; and b) reversing the causal pathway between the risk perception and behavioral response variables (i.e., personal factors behavioral responses) in Figure 5 without changing others . In the case of the first reversed model, the beha vioral response and risk perception variables have a significant and positive relationship at the p<.001. However, its model fits are drastically reduced compar ed to the research model and do not meet acceptable levels of model fit (e.g., RMSEA: 0.038 0.196, CFI: 0.970 0.417, TLI: 0.867 1.626, and SRMR: 0.010 . In terms of the second reve rsed model, while its model fit scores are only slightly weakened compared to the research model (e.g., RMSEA: 0.038 0.050, CFI: 0.970 0.949, TLI: 0.8 67 SRMR: 0.010 , the behavioral response variable do es not significantly affect the risk perception variable. Consequently, i t may be difficult to perfectly certify which pathway between the risk perception and behavioral response vari ables is right, based on the cross sectional survey data. Nevertheless, considering those results, the causal pathway from climate change risk perception to behavioral responses to mitigate climate change seems more convincing than the reversed pathway.

PAGE 106

92 Policy Implications The results of this national level analysis reveal that most personal factors significantly alleviate climate change directly or indirectly ( via climate chan ge risk perception ) . T here is a need for government efforts th at facilitate positive factors on climate change risk perception and individu mitigate the effects o f negative fac tors on them. R elevant pol icy implications are as follows: First, policies for helping individuals understand long term climate change experiences are necessary. Unlike short term experiences, long term climate change experiences, including the patterns o f extreme weather events, influence behavioral responses via climate change risk perception. Not only can miscommunication concerning climate change cause the general responses to address climate change. Short term experiences wo uld no longer be a suitable way of communicating with the general public regarding climate change in South Korea . Thus, policymakers should pay attention to the way in which they can help individuals perceive long term climate change ex periences as well as remind individuals of the importance of these experiences. Previous research projects provide some meaningful lessons . Rebetez (1996) emphasize s an appropriate form of data represent ation , such as the frequency calendar and cumulative frequency graph related to weather events . Thaker, Zhao, and Leiserowitz (2017) highlight the reality a nd accuracy of climate change in formation , including mitigation and adaptation strategies . Yu et al . (2013) indicate the importance of more current climate change information delivered timely. In the era of ICT , the timely provision of long term, up to date , and precise information related to climate change can be realized b ased on the establishment of an on line climate change information system . By the system, individuals can easily and quickly

PAGE 107

93 obtain climate change information and clearly and confidently perceive their long term climate change experiences. Second, systematic education on climate change needs to begin in earnest. A wareness of climate change, including its major cause and f ormidable outcomes, can lead to the increase of climate change risk perception , and this risk perception finally enhance s responses to curb climate change. Therefore, policymakers must seek to educate the general public about climate change. Yu et al . (2013) suggest the expansion of publicity to diffuse the knowledge of climate change. Bai and Liu (2013) advise the active use of low carbon education facilitating low carbon awareness and action skills of the general public . Berse (2017) proposes the educational system equipped with educational materials change. Policymakers should utilize diverse education methods (e.g., on line information system and educational broadcasting ), and those methods should target all citizens , regardless of their socio demographic characteristics . Particularly , given that age significantly and positively affect s behavioral responses to alleviate climate change, c limate change education included in the primary, sec ondary, or tertiary curriculums can be a proper way to expand the behavioral responses of the young er generation s . Third, governments should set a goal of finding a proper balance between the level of their efforts and that of individual efforts to address climate change. ongoing efforts to enhanc e climate policy are essential to minimize the adverse effects of climate change, not only can too much gove rnment driven efforts accompanied by huge investments inevitably increase the burden on government coffers, but they can also accelerate the contraction of individuals Also, t he general public might feel that governments and municipalities have to assume full responsibility. In fact,

PAGE 108

94 a ccording to the K water survey , the vast majority of the respondents think that the government and municipalities should play a big ger role in addressing climate change . 83 So, complementary limate policy have to be implemented . A variety of attractive financial incentives, such as tax reduction linked to less driving, subsidy and free parking service for electric cars, and electricity or water bill reduction based on an annual energy or water usage pattern , might be plausible strategies . Last , governments are asked to make diligent efforts to share their GHG reduction target and detailed plans to achieve it with individuals. The f amiliarity with the government plans es to mitigate climate change. But m ore importantly, as shown in Table 5 , policy related knowledge intensifies indirect behavioral responses to mitigate climate change, while climate change risk perception facilitates direct behavioral responses. This factor can be a great companion for risk perception in encouraging curb climate change. Further, s ome indirect responses, such as donation s to environmental organizations and voting based on environmental related election promises, involve political and social impacts and plac path of a virtuous circle. 84 G overnments should make reasonable goals related to the reduction of GHG emissions , unveil their achievement to the public in a transparent manner, and try to propagate the significant role These activities c an be effectively achieved with the implementation of the first two implications. 83 Concerning the question of who should play the most important role in addressing climate change, 81.3% of the respondents indicate the government (76.0%) or municipalities (5.3%), whereas 11.6% of them point out individuals. The remaining respondents (7.1%) select private and public cor porations, educational institutions, etc. 84 If individuals donate more money to environmental organizations and vote more enthusiastically for environment oriented politicians, these organizations will try to strengthen a national consensus about the serio usness of climate change, and policymakers will try to implement more effective climate policies . These efforts will attract more public attention and motivate more people to take behavioral responses against climate change.

PAGE 109

95 In brief , governments should try to enhance to address climate change by implement ing a variety of policies, such as providing proper climate change information by a Web based integrated system , diffusing climate change and policy related knowledge by various educational methods, and utilizing fina ncial incentives. A balanced set of on a particular policy. The thing policymakers always remember is not the implementation of those policies, but the iterative refinement of them. Limitations and Future Research Directions T he main limitation s of this empirical analysis are two fold. First, while this analysis focuses on risk perception, to address climate change also can be affected by d external contexts. For instance , behavioral responses to alleviate climate change are mostly in line with pro environmental behaviors (e.g., minimizing resou rce and energy consumption) practiced by the public for a long time. Those who have positive environmental value are likely to practice pro environmental actions ( Brandon & Lewis, 1999; Kollmuss & Agyeman, 2002) , regardless of the level of climate change risk perception. Tangible or intangible benefits, such as saving money and improving health, can be a significant motivation to engage in behavioral responses against climate change ( Bai & Liu, 2013; Whitmarsh, 2009a) . External contexts, such as the condit ion of public transportation and technological developments related to electric facilities and cars, may also cause actions that correspond to behavioral re sponses to address climate change. Second, instruments for measuring variables have room for improvement. Many personal factors used in this analysis, including climate change risk perception, are measured by a single item due to the limitation of utilizing existing data . Especially , s hort term experience

PAGE 110

96 variables do not reflect the frequenc ies and intensities of extreme weather events. Depending on single item measures is likely to weaken reliabilities ( Keller, Siegrist, & Gutscher, 2006). This analysis may have underestimated (or overestimated) the effects of those variables on climate change risk perception an d behavioral responses to mitigate climate change . T he findings of this national level analysis should be considered not as the last pitch , but a new momentum for unveiling this topic. Based on the limitations above, t he following future research directions are put forward. First , future research should try to investigate process es initiating behavioral responses to address climate change more precisely by utilizing a wider range of influential factors (e.g., risk perception, environmental propensity, and economic benefits) . The expansion of a research scope will provide a rich sour ce of policy alternatives aimed . A long with these efforts , researchers need to identify favorable circumstances and potential barriers to the implementation of those policy alternatives and expected outcomes . Second, to obtain more robus t results, future studies should be conducted based on much more refined and standardized survey instruments . Most previous research projects exploring similar topics have utilized cross sectional survey data measured by different questions and point scales . 85 The stepwise refinement and standardization of survey instruments would not only lead to more robust analyses in cross sectional studies , but also increase the likelihood of comparative and longitudinal studies. 86 85 In fact, experimental and lon gitudinal nature of research is better than non experimental and cross sectional studies in identifying causality ( Grothmann & Reusswig, 2006; Mase, Gramig, & Prokopy, 2017; Singleton & Straits, 2017; Thaker, Zhao, & Leiserowitz, 2017). However, when exploring this topic , the latter has been preferable to the former due to practical problems such as a lack of time, money, and instruments. 86 This direction somewhat echoes the argument of Marquart Pyatt et al . (2014). They emphasize the importance of des This would help increase the likelihood that future teams of social scientists utilize the same reliable, valid measures of climate change perceptions, ultimately improving the cohe rence in this literature, and accumulation of knowledge, over the long Marquart Pyatt et al ., 2014 , p. 255).

PAGE 111

97 Conclusion This national level analysis explores the causal relationship among personal factors, climate change risk perception, and behavioral responses to address (mitigate) climate change. SEM r egression analysis is conducted based on survey data of South Korea, and the results As predicted, climate change risk perception is significantly and positively correlated with behavioral responses to mitigate climate change. This evidence is consistent with the ar gument of many previous studies that climate change risk change. All the know ledge factors positively affect behavioral responses directly or indirectly via climate change risk perce ption . Especially, considering the analysis results of the sub models, policy related knowledge , promoting indirect behavioral responses such as donations to environmental organizations and voting based on environmental related election promises, shows the potential to complement climate change risk perception which mostly facilitates direct behavioral resp onses such as saving energy and utilizing renewable energy . While long term climate change exp eriences have a positive relationship with behavioral responses indirectly via climate change risk perception , contrary to expectation, short term (extreme w eather event) experiences fai l to intensify both risk perception and behavioral responses . Also, only the assessment of drought policy , exce pt that of flood policy, negatively influence s behavioral responses mediated by climate change risk perception. Among control variables, gender and age have impacts on risk perception or behavioral responses associated with climate change. Despite some lim itations, t here is no doubt that this national level analysis conveys significant information on the causal relationship among personal factors, risk perception, and behavioral responses associated with climate change .

PAGE 112

98 The findings of this analysis leave some significant policy implication s . Simply, p long term climate change experiences, how to initiate systematic education on climate change, how to share their GHG reduction plan with individuals , and how to keep a balance Developing a Web based integrated information system and utilizing various educational programs an d financial incentives might be practical methods to settle these pending questions. Besides climate change risk perception, other s , financial benefits, and social and technological contexts , related to climate change. When exploring this topic in the future , research ers need to expand a rang e of influential factors beyond risk perception. Also, they have to keep in mind that continu ed efforts to refine and standardize survey instruments related to climate change risk perception and c ould lead to more robust and fruitful results.

PAGE 113

99 CHAPTER V LOCAL LEVEL ANALYSIS South Korea has been regarded as a country with a water shortage by water related international organizations. 87 Especially, South Chungcheong Province, situated in the west of South Korea, has repeatedly experienced droughts since 2010 (Ministry of Infrastructure and Transport (MOLIT) & K water, 2017). B ased on survey data of individuals living in South Chun gcheong Province , this chapter investigate s the causal relat ionship among personal factors, climate change ( drought ) risk percepti on, and behavioral responses to adapt to climate change ( drought ) . T he results of the analysis will be helpful to South Chungcheong Province in overcoming droughts and shaping effective and efficient climate policies. A lso , given that data were collected f r o m a face to face survey during the 2015 drought period , the findings might help both policymakers in many drought prone areas and scholars across the world better understand Research Context s Water M anagement and Climate C hange Water has played an important role in human history. The major ancient civilizations of the world were developed based on efforts to manage water, such as the Nile, Indus, Yellow, Tigris, and Euphrates Rivers ( LaMoreaux, 2001) . Since the Industrial Revolution, water management has contributed to successful industrialization and urbanization i n many countries through providing domestic and industrial water (K water, 2016 a ), generating hydropower 87 The amount of available water per capita (index) is commonly used by several organizations, such as the United Nations Environmental Program and the Population Action International, for determining the conditions of each g to the data released by the Population Action International in 2003, the index of South Korea was 1,453 cubic meters. It was ranked 129 among 153 nations. This organization categorized South Korea within the group of water stressed countries (i.e., the i ndex is less than 1,700 cubic meters) (MOLIT & K water, 2017) .

PAGE 114

100 ( Tvedt, 2010) , transporting heavy materials and products cheaply ( Aldcroft, 1983) , and securing stable urban environment (Novotny, Ahern, & Brown, 2010) . However, drastic climate change in the 21st century has made it difficult to manage water quantity as well as quality and requires more water related investments ( Bates et al ., 2008) . Many nations across the world are experiencing climate ch ange and South Korea is no exception. Water M anagement in South Korea Since the 1960s, the South Korean government has initiated full scale water resources development based on long term plans, such as the 1st Comprehensive Water Resources Development (10 years) the Comprehensive Four Major River Basin (K water, 2016 a ) . Many hydroelectric dams (e.g., Uiam (1967: The time of completion) and Paldang (1974)) and multi purpose dams (e.g., Seomjin River (1965), Soyanggang (1973), and Andong (1977)), were constructed to secure water resources and energy ( MOLIT & K water, 2017) . The construction of wide area waterworks, which can provide raw or pu rified water to two or more local governments (Min istry of Government Legislation , 2016 ) , also started nationwide in the 1970s to supply domestic and industrial water. As a result, in 2000, the percentage of households with direct, potable water reached 87.1 . After 2000 , water policies were transformed from securing water quantity to restoring water quality and ecosystems based on the 4th (K water, 2016 a ) . T he increasing construction cost of dams and public c oncerns about environmental damage have made the government focus on the development of new water sources, such as small dam construction and desalination, instead of the development of large scale dams (MOLIT & K water, 2017) .

PAGE 115

101 According to the United Nations Water (2006), rapid urbanization, population concentration, and droughts triggered by clim ate change have exacerbated global water scarcity. South Korea also has not directly suffered from less precipitation, but high density of population, sharp f luctuations of seasonal precipitation, and the intensification of climate change make it difficult to manage water resources. For example, the annual precipitati on per capita is just 2,546 cubic meters , one sixth of the world average, whereas the average a nnual precipitation is 1,300 (MOLIT & K water, 2017) . Moreover, since the 2000s, severe regional droughts are getting more frequent in South Korea. Water Condition in South Chungcheong Province South Chungcheong Province , abbreviated as Chungnam , is located in the west of South Korea. T he Boryeong dam, the main water source for several cities in Chungnam , was constructed in 1998. Since then, it has supplied water regularly to Chungnam for over 15 years. The total water storage capacity of the dam is 116.6 million cubic meters , including flood control storage capacity (i.e., 10 million cubic meters ). The water supply capacity is 106.6 million cubic meters (i.e., domestic and in dustrial wate r 90.6 million , agr icultural water 4.7 million , and stream mai ntenance water 11.3 million cubic meters ) (K water, 2013) . In 2015, the Boryeong dam was nearly drained due to a prolonged drought in the region. The amount of water i n the Boryeong dam was just around twenty two percent of its capacit y (Kim, Chang, & Won, 2015). Chungnam inevitably planned to cut down water supply in eight cities (e.g., Boryeong and Seosan cities) ( Song, 2015) . MOLIT and K water , who were in charge of water (quantity) management in South Korea, considered several measures to tackle the water shortage in Chungnam . Urgently, they construct ed a pipeline from one of the sixteen dammed pools (reservoirs) on the four major rivers (i.e., the Han, Geum, Yeongsan, and Nakdong Rivers) ,

PAGE 116

102 called the Baekje bo , to the Boryeong dam ( MOLIT, 2015) . Unfortunately, Chungnam shortage in 2015 was not the last. As of June 5, 2017, the national average of accumulated precipitation was just 166.6 millimete rs, down forty seven percent from the previous year. It marked the second lowest record since the government started to make weather forecasts in 1973 ( Kim, 2017) . The average amount of water in reservoirs also fell under fifty percent ( Kim, 2017) . In 2017, once again, Chungnam suffered from water scarcity ( Choi, 2017) . W ater scarcity occurs for two main reasons: a) the excessive use of water or the inefficient operation of water sources and facilities (i.e., man made water scarcity); and b) the lack of wa ter quantity brought by drought (i.e., natural water scarcity) ( Pereira, Cordery, & Iacovides, 2009) . According to Collier and Webb (2002), climate change has intensified extreme cycles involving severe droughts and deadly floods. There is no evidenc e that K water, which is responsible for the management of the Boryeong dam, failed to control the dam water level, and citizens and industries in Chungnam excessively used water. Namely , the lack of water quantity brought by drought is the cause of Chungnam To solve this problem , securing more water in advance or saving water during drought (or both of them) , can be a possible solution . While the former (e.g., building new dams and desalination facilities) typically requires time an d money and is likely to come up against sturdy opposition by citizens and environmental organizations, the latter saving effort) can be an inexpensive and less time consuming settlement with no severe resistance. Overall, it is increasingly difficult to manage water due to climate change and there is a high probability that drought will continue to occur in Chungnam . In this regard, the role of individuals in addressing drou ght will be much more important, and t his analysis, could offer practical and econo mical lessons to adapt to natural water scarcity in Chungnam .

PAGE 117

103 Research Methods Data Collection Research tool . This analysis is based on survey data of local citizens in Chungnam . In late 2015, the K water i nstitute conducted a survey in north west Chungnam to investigate water consumption pattern and water policy preference under the drought in the second half of 2015. North west Chungnam consists of three cities (i.e., Dangjin, Boryeong, and Seosan) and five counties (i.e., Seocheon, Yesan, Cheongyang, Taean, and Hongseong ) . A regionally representative quota sample of north west Chungnam population aged 19 years and old er was surveyed face to face i n their own homes between 10 and 23 th December, 2015 . 88 Survey design and representativeness . Similar to the questionnaire used in the national level survey , r esearchers at the research center for water policy and e conomy in the K water institute developed a questionnaire . Before conducting the survey, they evaluate d and improve d the design of the questions based on expert reviews of the survey questionnaire . The questionnaire includes key questions associated with the following topic s: a) drought risk perception (e.g., the seriousness of drought) ; b) self vulnerability perception related to limited water supply (e.g., tolerance against limited water supply situations ) ; c ) experiences of water problems during drought (e.g., experiences of water supply restriction and low water pressure) ; and d ) behavioral responses to adapt to drought (e.g., daily actions and device utilization related to water saving) . Most questions are closed ended , mainly in the form of a four point scale . Appendix F lays out the key questions of the questionnaire. 88 This study views that most individuals living in Chungnam associate drought with climate change. For instance, in 2016, the Chungnam institute surveyed 1,000 local citizens throughout Chungna m. More than ninety five percent of the respondents believed that climate change ( moderately , seriously , or very seriously ) intensifies drought and increases its frequency in their province (Chungnam institute, 2016). Also, according to the survey data of the national level analysis, all respondents who reside in Chungnam (N=51) answered that the danger of drought is either somewhat or definitely increasing in the last ten years in relation to climate change.

PAGE 118

104 Table 6 2015 s urvey r demographic characteristics Classification Respondents Population in n orth west area of Chungnam Num. Per. Num. Per. Gender Male 156 30 .3% 394,464 50.9% Female 359 69.7% 380,802 49.1% Age (0 18) (130,811) 19 39 68 13.2% 182,377 28.3% 40 49 93 18.0% 115,738 18.0% 50 59 160 31.1% 125,546 19.5% 60 and above 193 37.5% 220,794 34.3% No answer 1 0.2% Education Under middle school 191 37.1% No High school 261 50.7% a vailable Undergraduate and above 61 11.8% d ata No answer 2 0.4% Income (Average Monthly Income before Tax) I
PAGE 119

105 Representativeness . The face to face survey, based on a structured questionnaire, was conducted by fully trained surveyors who belong to the Gallup Korea (http://www.gallup.co.kr/) and took approximately one hour on average to complete. Individuals who reside in north west Chungnam , around 770,000 as of 2016, was the target population (Statistics Korea, 2017b) . Total 515 respondents were selected randomly from a stratified sample of the eight areas. The socio demographic characteristics of the respondents, along with those of north west Chungnam population, are presented in Table 6 . The proportions of female (69.7%) and 50s age (31.1%) in the respondents appear to be somewhat higher than those of female (49.1%) and 50s age (19.5%) in north west Chungnam population. Regarding the patterns of water consumption in South Korea, women play a key role in domestic water use, as they typically bear primary responsibility for water intensive activities such as washing dishes and doing laundry. Even though these two groups are over represented in the dataset, this study views that it is not serious ly problematic because they are more likely to be aware of water issues in the household and play a significant role in deciding water consumption patterns of the household. Variable Description Mediating and d ependent variable s . Drought risk perception and behavioral responses to address drought are utilized as a mediating (both dependent and explanatory) variable and a dependent variable , respectively . Drought risk perception . Similar to climate change risk perceptio n which is measured by the perceived seriousness of climate change , r espondents were asked to indicate their level of seriousness of drought in Chungnam o n a four point scale, where 1 was very serious and 4 was not serious . This analysis reverses the scale of the question so that higher values indicate greater perception of the seriousness of the problem.

PAGE 120

106 Behavioral responses to address drought . From an individual perspective, actions to save water (reduce water consumption) are practicab le behavioral efforts to adapt to drought, which is viewed as a water shortage due to less precipitation than normal. 89 Two types of behavioral response s , daily action and device utilization , are utilized as dependent variables. In the case of the daily action, a total of seven items are used to create a holistic assessment of strongly , 2 = moderately , 3 = hardly , and 4 = not , the respondents were asked to indicate the level of how much they were committed to practicing the following water saving behaviors than usual: a) using a glass with water when brushing my teeth; b) using a wash basin with water when washing my face; c) larger quantity; f) using a washing up bowl when washing dishes; and g) inspecting a leakage of water. This study reversely codes the scale of the questions, wh ere 0 = not , 1 = hardly , 2 = moderately , and 3 = strongly , so that higher values indicate more efforts on daily actions to save water. The index of daily action is created by summing each value of the seven questions 90 The range of this index is from 0 to 21. In terms of the device utilization, on a dichotomous scale, where 1 = yes and 2 = no , the respondents were asked to indicate whether they were using the following water saving devices: a) a water saving faucet; b) a water sa ving toilet bowl; and c) a water saving shower. The scale of the questions are recoded, where 0 = no and 1 = yes . Like the index of daily action, each value of the three questions is 89 In fact, Appendix A does not clearly provide the conceptual map for the local level analysis. Nevertheless, it, 90 Principal component factor analysis indicates that the number of retained factor is one . While the first factor has an Eigenvalue of 3.61 which is 52% of the total possible of 7, the second factor has an Eigenvalue of just 0.77. All correlation values are greater than 0.3. The overall value of KMO measure, 0.8598, is placed in the meritoriou s sum, those seven questions are suitable factors for establishing the index.

PAGE 121

107 185). 91 In other words , the higher the index is, the more water saving devices a respondent uses. Independent variables . One experience variable , three knowledge variables , and two policy assessment variables, along with self vulnerability perception related to limited water supply , are employed as independent variable s . Experience of water problems . It is not easy for the general public to know that they are experiencing drought because of the difficulty in recognizing w hen it starts and ends which Wilhite (2000) introduces as one of the characteristics of drought . Instead, this study argues that the restriction of residential water supply due to a water shortage can be the most obvious experience of drought to them. The respondents were asked whether or not they experienced any water supply problems, such as limited water supply, low pressure, etc., during the drought . e xperience of water problems is coded as a dichotomous variable, wher e no = 0 and yes = 1. Knowledge . Two general knowledge variables (i.e., awareness of water quantity condition at the national level and the exposure to water saving information ) and one specific knowledge variable (i.e., the familiarity with a subsidy policy related to drought ) were measured. 92 First, concerning water quantity condition in South Korea , respondents were asked to select their answers among four items, where 1 = very abundant , 2 = abundant , 3 = scarce , 91 The critical point of Eigenvalue is 1 in principal com ponent factor analysis. The Eigenvalues of the first factor and second factor are 2.21 (74% of the total possible of 3) and 0.51, respectively. There is no correlation value ificant at the p<.001 level, and the items are adequate for generating the index. 92 Unlike climate change, the general public is quite familiar wit h drought itself and its causes (e.g., less consumption during drought is one of the important strategies to governments in addressing drought. Therefore, this study believes that the above mentioned variables might be appropriate indicators for drought knowledge.

PAGE 122

108 and 4 = very scarce . Second , the exposure to water saving information through a campaign, publicity, or education is simply coded as a dichotomous variable where no = 0 and yes = 1 . familiarity with a gove rnment subsidy policy associated with water saving during a drought period was measured on a four point scale, where 1 was precisely and 4 was not . This study reverses the scale of the question , where 1 = not , 2 = hardly , 3 = moderately , and 4 = precisely . Therefore, h igher values point out greater familiarity with the government subsidy policy related to water saving during a drought period . Policy assessment . Core policies influencing water management during drought are water supply and water infrastructure policies (K water, 2016 b ) . I n the case of water supply policy, the respondents were asked to answer their level of evaluation for in convenience in using water in Chungnam on a four point scale, where 1 = very serious , 2 = moderately serious , 3 = hardly serious , and 4 = not serious . In terms of water infrastructure policy , the respondents were asked to answer their level of eval uation for water infrastru cture in Chungnam on a four point scale, where 1 = very good , 2 = good , 3 = poor , and 4 = very poor , and t heir answers are reverse coded. Overall , higher values point out higher satisfaction with water management policies. Self vulnerability perception related to limited water supply . T his analysis utilizes the perception of self vulnerab ility to the restriction of water supply as an additional ind ependent variable. When facing a prolonged drought, there is a high possibility of cutting off water supply intermittently, and the execution of limited water supply tends to cause both inconvenience for individuals and losses to industries. Especially, people who themselves feel highly vulnerable to limited water supply are likely to suffer more inconvenience from actual limited water supply than people who themselves do not, and the former might want to avoid the restriction of water supply more than the latter. Generally, the public notices that not only can water saving efforts

PAGE 123

109 by individuals and industries delay the timing of limited water supply, but also shorten the hours of restricting water supply. Thus , this study views that higher se lf vulnerability perception has a 93 limited water supply. On a five point scale, wh ere 1 = from one to three days , 2 = from four to seven days , 3 = from eight days to less than a month , 4 = from one month to less than three months , and 5 = more than three months , the respondents were asked to indicate the level of how long they can toler ate the following limited water supply situations (per day): a) from three to less than six hours; b) from six to less than twelve hours; c) from twelve to less than eighteen hours; and d) from eighteen to less than twenty four hours. Similar to drought ri sk perception, the scale of the questions are reversely coded, where 1 = more than three months , 2 = from one month to less than three months , 3 = from eight days to less than a month , 4 = from four to seven days , and 5 = from one to three days . Based on the summation of each value of the questions, the index of self 94 The range of 93 This variable was recommended by K water researchers, who have explored drought for a long time and conducted the local level survey, as a potential factor influencing individu saving behaviors during drought. This study tested several pilot models with self vulnerability perception as a mediating variable by adding pathways from other (independent and control) variables. Their model fits are poor, whereas few significant pathways are identified. For example, in the pilot model assuming that s elf vulnerability perception variable is affected by water problem experience, the exposure to water saving information, and socio demographic variables, 30.945 (df = 9, p<.001), RMSEA 0.070, CFI 0.862, TLI 0.447, and SRMR 0.022). Also, the literature does not provide enough evidence that this variable can act as a mediating variable in the research model. Considering these results and the core of this stu responses, this analysis employs self vulnerability perception as an independent variable. According to the regression analysis results, utilizing self vulnerability per ception as an independent variable does not significantly change the results of the relationships among other variables and improves model fits. 94 Given the results of principal component factor analysis, the number of retained factor is one. Specifically, the Eigenvalue of the first factor is 3.08, and this is 77% of the total possible of 4. By comparison, that of the second factor is just 0.54. There is no correlation value which is 0.3 or less than 0.3. The overall value of KMO measure is 0.7806, and it significant at the p<.001 level. Overall, those four questions are suitable factors for creating the index.

PAGE 124

110 this index is from 4 to 20. In other words, t he higher the index is, the shorter period of limited water supply a respondent can tolerate (i.e., the higher levels of self vulnerability perception a respondent has). Control variables . To isolate the influence of personal factors on drought risk perception and behavioral responses to adapt to drought, this study utilizes five socio demographic variables. E ach variable is coded in a similar way that the national level analysis does. First, gender is measured as a dichotomous variable , where 0 = male and 1 = female . Second, age is expressed in years . Third, th e level of education is coded on a three point scale, where 1 = under middle school , 2 = high school , and 3 = undergraduate and above . Fourth, the level of income (i.e., average monthly income before tax last year) is coded on a seven point scale , range from 1 for those earning less than KRW 1million (around US$ 909) to 7 for those earning over KRW 6million (around US$ 5,455) . Lastly, residential area is coded as a dichotomous variable, where 0 = urban area (city ) and 1 = rural area (county ) . Data Analysis Considering the structural relationship among personal factors, risk perception, and behavioral responses related to drought, the analysis plan of this chapter is along th e same lines as that of the previous national level chapter . SEM is employed to test hypotheses. This empirical analysis is performed based on the following steps. F irst, socio demographic characteristics and descriptive results are presented. Subsequently , biv ariate correlations between the dependent variable s and other variables are measured . Finally , SEM regression analysis is conducted by utiliz ing STATA , following the evaluation of model fit . Pathways in the local level research model is described in Figure 9. Prior to this research model, some pilot models, adding pathways among p ersonal factor variables, are analyzed and provide mixed results with poor

PAGE 125

111 model fits. 95 Following the national level analysis, the relationship among the personal factors is not covered in the local level analysis. Figure 9 . Pathways in the local level research model Analysis Results Socio demographic Characteristics The respondents range in age from 20 to 83 year s (M = 54.38, SD = 12.12) and 69 .7% are female s . The majority of the respondent s are equal to a high school graduate in scholastic achievement , and or graduate degree . About 25.6 % earn over KRW 3million (US$ 2,727) a month. 95 For example, one of the pilot models includes the pathways between water problem experience and policy assessment and between general knowledge and policy related knowledge. According to the regression analysis results, while the assessment of water supply is significantly and negatively affected by w ater problem experience (p<.001), the assessment of water infrastructure is not. Also, while the exposure to water saving information significantly and positively affects the familiarity with a subsidy policy (p<.001), awareness of water quantity condition 0.085, CFI 0.605, and TLI 0.298).

PAGE 126

112 Descriptive Results Most of the respondents are aware of the inferior condition of water quantity at the national level (M = 2.83, SD = 0.46 ) . N early 80% of the respondents indicate that water quantity in South Korea is either scarce (76 .7 %) or very scarce (3.1%). More than half (55.9%) have been exposed to a campaign, publicity, or education related to water saving (M = 0.56, SD = 0.50). When asked to consider the seriousness of drought in Chungnam, approximately three quarters of those surveyed perceive that it is either serious (60.8 %) or very serious (13.4 %) (M = 2.88, SD = 0.61) . Concerning personal experiences, 13.6% of the respondents have experiences of water supply problems, such as limited water supply and low pressure, during this drought. In terms of the 82.3% and 61.1 % of those surveyed positively (either good or very good ) assess the water supply and infrastructure p olicies , respectively , whereas 17 .7% and 38.9 % of those surveyed negatively (either poor or very poor ) evaluate them, respectively. The proportion of the respo ndents who are familiar with the government subsidy policy related to water saving ( either moderately or precisely ) is a little bit less than that of the respondents who are not ( either hardly or not ) ( M = 2.12, SD = 0.99 ). Lastly, as shown in Figure 10 , the majority of the respondents answer that they are either strongly or moderately practicing the following water saving behaviors than usual : a) using a glass with water when brushing my teeth (M = 3.41, SD = 0.74); b) using a wash basin with water when w ashing my face (M = 3.30, SD = 0.81); c) shortening shower time (M = 3.36, SD = a larger quantity (M = 3.49, SD = 0.67); f) using a washing up bowl when wash ing dishes (M = 3.15, SD = 0.86); and g) inspecting a leakage of water (M = 3.31, SD = 0.66 ) . All items are

PAGE 127

11 3 practiced by more than 80% of those surveyed. F our items among them, such as the third, fourth, fifth, and seventh , are practiced by over 90% of the respondents. Figure 10 . R esults of seven water saving efforts than usual Bivariate Cor relation Coefficients among Variables The inter correlations of the key variables in the research mode l are shown in Table 7 . In the case of the relationship between drought risk perception and independent variables (Table 7 .1) , all relationships between them show the hypothesi zed directions (i.e., one experience and two knowledge factors positive; two policy assessment factors negative). The f our independent variables except the assessment of water supply show a significant correlation with drought risk perception . Amo ng the socio demographic variables, gender and income have a positive relationship at the p< .01 level and a negative relationship at the p<.05 level, respectively. In terms of the correlation between beha vioral responses against drought and other variables , including drought risk perception (Table 7 .2) , as expected, the familiarity with a subsidy policy associated with water saving is significantly and positively correlated w ith both daily action and device utilization responses . Meanwhile , the exposure to water saving information show s a significant positive correlation only with daily action responses (p<.001).

PAGE 128

114

PAGE 129

115 Surprisingly, drought risk perception and water problem experience s do not have a significant relationship with both daily action and device utilization responses . While self vulnerability perception rela ted to limited water supply has a significant correlation with both behavioral responses Regarding socio demographic variables, age, education, income, and area have a si gnificant relationship with daily action, device utilization, or both responses . Gender does not have any significant effects on both behavioral responses. M odel Fit The values of model fit indices for the = 5, p>.05), RMSEA = .047(90% CI), SRMR = .011, CFI = .959, and TLI = .793. Also, those of model fit indices for the device utilization model are as follows : RMSEA = .000(90% CI), SRMR = .007, CFI = 1.000, and TLI = 1.051. Considering the conventional cutoff values of fit indices (see F ootnote 77 ), the ir model fit s are acceptable . Regression Analysis T o identify the relationship among personal factor s , risk perception, and behavioral responses re lated to dr ought, this chapter employs SEM regression analysis utilizing maximum likelihood estimation with robust standard errors. 96 Those things that personal factors affect drought risk perception ( H1 , H2 , and H4 ) and some personal factors and drought risk perception (and self vulnerability perception) influence behavioral responses to adapt to drought ( H1 , H2, 96 This local level analysis also conducts univariate and multivariate normality tests for variables in the research models. Similar to the national level analysis, skewness and kurtosis) test indicate that multivariate n ormality assumption is violated, whereas all univariate skewness and kurtosis of each variable are not located within the range of non normality concern (i.e., |skewness| > 2, |Kurtosis| > 7) ( Chou & Bentler, 1995; West, Finch, & Curran, 1995). Considering the identified approaches to dealing with non normality, Appendix G provides the results of SEM regression analyses with different estimation methods (or options). Consequently, directions and significance of coefficients are almost identical. Thus, the r esults of ML with robust Additionally, multicollinearity, influential outliers, and heteroscedasticity are not detected in both research models.

PAGE 130

116 H3 , and H5 ) were hypoth esized in the literature review chapter [III] . Table 8 and Figure 11 lay out the results of the SEM regression analysis and significant pathways in the two models, respectively. Also, Appendix H shows direct, indirect, and total effects of independent variables on behavioral responses to address drought. Table 8 . R esults of SEM regression analyses about the relatio nship among personal factor s , risk perception, and behavioral responses related to drought Dependent Variables Independent Variables Daily action model Device utilization model Coeff. Robust Std. Err. Coeff. Robust Std. Err. Drought risk perception Water problem experience .094 .076 .105 .076 Awareness of water quantity condition .206*** .058 .207*** .058 Exposure to water saving information .198*** .056 .192** .056 Assessment of water supply policy .007 .059 .002 .059 Assessment of water infrastructure policy .186*** .051 .191*** .051 Gender (female) .124* .058 .127* .058 Age .001 .003 .001 .003 Education .022 .056 .017 .056 Income .047* .022 .048* .022 Area .045 .058 .043 .058 Constant 2.787 2.770 Behavioral responses Drought risk perception .040 .250 .053 .057 Self vulnerability perception .101 .051 .078*** .014 Water problem experience .407 .484 .092 .103 Ex posure to water saving information 1.049** .372 .013 .084 Familiarity with a subsidy policy .398* .161 .119** .038 Gender (female) .568 .339 .037 .086 Age .070*** .015 .001 .004 Education .301 .324 .266** .085 Income .206 .115 .141*** .037 Area .931** .337 .551*** .083 Constant 11.834 1.699 n 501 503 Note : Fourteen respondents who did not answer at least one question out of the seven questions related to daily action and twelve respondents who did not answer at least one question out of the three questions associated with device utilization are excluded from the SEM regression analyses of the first model and second model, respectively. *p<.05, ** p<.01, ***p<.001 (two tailed test)

PAGE 131

117 *p<.05, ** p<.01, ***p<.001 (two tailed test) Figure 11 . Significant pathways in the daily action and device utilization models Drought risk perception . Three independent and two socio demographic variables are significantly correlated with drought risk perception. 97 Consequently, the results of the SEM 97 Regard ing drought risk perception, the first and second models show the slightly different coefficients of independent variables due to the different numbers of respondents. However, there is no significant difference. This study explains the results of the anal ysis based on the daily action model.

PAGE 132

118 re gression analysis support Hypothesis 2 and partially verify Hypothesis 4 . They fail to substantiate Hypothesis 1 . Experience . This study finds no significant relationship between water problem experiences, such as limited water supply and low pressure, and drought risk perception. This result indicates that t here is no evidence that the respondents with water problem experiences are likely to have higher levels of drought risk perception. It is reminiscent of the national level od and drought) do not have a significant impact on climate change risk perception. Again, exploring this relationship more in depth is recommended . General knowledge . As predicted, drought risk perception is positively affected by both awareness of water quantity condition (b = .206, p<.001) and the exposure to water saving information (b = .198, p<.0 0 1). The respondents who are more aware of water quantity condition s in South Korea or have been exposed to water saving information tend to have higher levels of drought risk perception. This result suppor ts the findings of the national level analysis as well as existing literature (e.g., Malka, Krosnick, & Langer, 2009; Shi et al . , 2016 , Wildavsky & Dake, 1990). Policy assessment . While t he assessment of water infrastructure policy has a significant negative relationship with drought risk perception , the assessment of water supply policy has no impact on drought risk perception. The respondents who have lower satisfaction with water infrastructure are likely to have higher levels of drought risk per ception. T his finding, along with that of the national level analysis, does not fully verify the hypothesized negative relationship between the performance of climate policy and climate change risk perception of the general public.

PAGE 133

119 Socio demographic variables . Gender and income significantly influence drought risk perception. While gender is positively correlated with drought risk perception (b = .124, p<.05) , income has a negative relationship with drought risk perception . The se results indicate that t he female respondents and low income respondents tend to be more concerned about drought than the male respondents and high income respondents, respectively. 98 Behavioral responses to address drought . Three independent and four socio demographic variables directly influence daily action, device utilization, or both responses. Considering the findings of the SEM r egression analysi s, while Hypothesis 2 and 3 are supported , Hypothesis 1 and 5 are not verified. Drought risk perception . A lthough this study hypothesizes that drought risk perception has a positive relationship with behavioral responses t o address drought, no evidence of a significant relationship between drought risk perception and behavioral respon ses is found in both the daily action and device utilization models. Some existing research (e.g., Bubeck, Botzen, & Aerts, 2012; Wachinger et al ., 2013) tries to explain the discrepancy (or the weak relationship) between risk perception and their behavioral responses to addr ess risks . For instance, Wachinger et al . (2013) report that the perceived benefits outweighing the potential adverse impacts, strong trust in the effectiveness of governance structures, and a lack of the situation that high risk perception does not lead to appropriate preparedness measures (p. 1054). B ased on the reviews of flood related empirical literature (e.g., Botzen, Aerts, & van den Bergh, 2009; Grothmann & 98 The positive impact of gender on drought risk perception is in accordance with the results of the national level analysis (see p. 83). Meanwhile, van der Linden (2017) states that while the negative relationship between income (and resources) and the self evaluation of vulnerability to the impacts of climate change has been hypothesized, empirical evidence provided by previous research projects are quite inconsis tent. The impact of income on drought risk perception in this analysis is also not consistent with the results of the national level analysis (see p. 83) .

PAGE 134

120 Reusswig, 2006; Siegrist & Gutscher, 2 006) , Bubeck, Botzen, and Aerts (2012) conclude that several factors, such as government compensation for damages, a lack of the sense of responsibility, fatalism, and wishful thinking , have the potential to inhibit behavioral responses of individuals having high risk perception against floods . 99 Their argument points to the potential need to account for values or worldviews in studying these behavioral questions, which was previously discussed in the final section of Chapter III. Overall, the explanation for this unexpected relationship is still unclear . This issue will be re assessed in the next chapter. E xperience . N o significant effect of water problem experience s on behavioral responses to address drought is detected in the models. The behavioral responses of the respondents with water problem experience are not s ignificantly different from those of the respondents with no water problem experience. This result suggests that forceful limited water supply or low water pressure time voluntary water saving during a drought period. Policy related and general kno wledge . The f amiliarity with a subsidy policy related to water saving during drought shows a significant positive correlation with both daily action (b = . 398, p<.05 ) and device utilization (b = . 119, p<.01 ) responses to save water . The respondents who a re more familiar with the subsi dy policy are not only more likely to change their daily actions, but also to utilize devices to save water. T his finding corroborates evidence about the influence of climate policy related knowledge on behavioral responses to address climate change in the national level analysis . 100 Meanwhile , the exposure to water saving information influences 99 model and Footnote 46 (see p. 44) help understand fatalism and wishful thinking . 100 It might be true that this variable is intertwined with a financial benefit, which is regarded as one of the potential Bai & Liu, 2013; Whitmarsh, 2009a).

PAGE 135

121 only daily action responses (b = 1.049, p<.01) . The respondents who have been exposed to water saving information are more likely to change their daily actions to save water, but they do not tend to use more water saving devices. Self vulnerability perception related to limited water supply . The correlation between self vulnerability perception and behavioral respons e variables is statistically significant in the device utilization model (p<.001). 101 However, contrary to expectations, it has a negative direction. The respondents who feel more vulnerable to limited water supply due to drought are less likely to practice device utilization to save water in the household. Some possibilities exist in inte rpreting these results. They may lack capacity or self efficacy, meaning they are not likely to perceive their actions as mattering. Also, the inconvenience of reducing water consum ption is greater than that of the respo ndents who feel less vulnerable. The refore, while the former may be reluctant to bear current high costs (i.e., inconvenience triggered by reducing water consumption) for uncer tain future limited water supplies , the latter may be willing to pay relatively low cost s . T his dissertation will return to t hese interpretation s in the expert level analysis . Socio demographic variables . Age (b = .070, p<.001) and area (b = .931, p<.01) are positively correlated with daily action responses . The elder respondents and those that live in count ies are more likely to change their daily actions to sa ve water than the younger respondents and those that live in cities , respectively. Education (b = .266, p<.01) , income p<.001) , and area have a significant relationship wit h device utilization responses . While education has a positive relationship with device utilization responses, income 101 When using bootstrapping or ADF, the relationship between self vulnerability perception and behavioral responses in the daily action model also becomes statistically significant (p<.05) (see Appendix G).

PAGE 136

122 and area have a negative relationship with them . The more highly educated respondents are more likely to use water saving devices in their household s than respondents with lower levels of education . 102 The high i ncome respond ents and those that live in counties are less likely to utilize water saving devi ces in the ir household s than low i ncom e respond ents and those that live in cities , respectively. Alternative Model Specifications Relationship between s ocio demographic characteristics and personal factors . Like the national level analysis, it is possible that socio demographic characteristic s directly influence the personal factors modeled in this research . Appendix E provides the results of regression analyses about the relationship between each personal factor variable and five socio demographic characteristic variables in the research model. Even though the se results identify some significant relationships between them , this analysis also excludes these relationships in the research model due to several theoretical and practical problems presented in the natio nal level analysis (see p . 90 ). Reverse pathway between risk perception and beh a vi o ral responses . T he reverse relationship between risk perception and behavioral responses associated with drought is also checked . It is tested in the same ways as the national level a nalysis: a) switching the places of the risk perception and behavioral response (daily action) variables in Fi gure 9 ; and b) reversing the causal pathway between the risk perception and behavioral response (daily action) variables in Figure 9 . The relationship between the behavioral response and risk perception variables is not statistically significant in bot h the first and second reversed models. The model fits of the first 102 This result is in line with the findings of previous studies (e.g., Leiserowitz, 2006; Patchen, 2006 ) that those with better education are more likely to take pro environmental actions than less educated people.

PAGE 137

123 one are drastically aggravated (e.g., RMSEA: 0.047 0.123, CFI: 0 .959 0.714, TLI: 0.793 0.431, and SRMR: 0.011 compared to the research model and are not acceptable. T hose of the second one are also slightly weakened compared to the research model (e.g., RMSEA: 0.047 0.051, CFI: 0.959 0.951, TLI: 0.793 SRMR: 0.011 0.011) . T hese results imply that the causal pathway from drought risk perception to behavioral responses to drought is more persuasive than the reversed pathway. Policy I mplications efforts to save water can be a great help to 103 However, in reality, individuals are less likely to recognize the need for water saving and to take actions to save water until governments announce that they are experien cing dro ught. As alluded to previously , drought is more difficult to cope with than flood s due to its peculiar characteristics . It is almost impossible for governments to accurately catch the starting point of drought and to effectively address severe drought without the cooperation of individuals. Also, it is not easy for individuals to realize the neces sity of their behavioral responses to save water and to voluntarily take action of drought. Therefore , to reduce the adverse effec behavioral responses to save water , and individuals need governments to help them understand the emergence of drought, its expected severity , and possible response methods. The results of this loca l level analys is convey several policy implication s to policyma kers . First, instead of depending on drought risk perception, active communication to facilitate While some personal 103 related a ctivities, such as the development of new water sources (e.g., dams, groundwater, and seawater desalination), the operation of multi purpose dams (e.g., discharge timing and volume and hydroelectric power generation), and the distribution of water (e.g., d omestic, industrial, and agricultural water), can be optimized based on self controlled water consumption of individuals.

PAGE 138

124 factors (e.g. , awareness of water quantity condition s and the assessment of water infrastruc ture policy) significantly affect drought risk perception , their impacts are not carried on to In other words , a t raditional drought strategy is for governments to stimulate drought ris k perception by transmitting simple information about local water (quantity) conditi on s to individuals , which presumably then leads . This, however, is not a reliable approach for addressing drought . r management during drought has a positive impact on public responses ( Whitmar sh, 2008). Two way communication (e.g., between governments and citizens and between citizens and citizens) will be another good alternative. 104 Lorenzoni and Pidgeon (2006) highlight that individuals are likely to take behavioral responses to address climate change when they 105 Second, the provision of more specific and practic al wisdom about water saving is required to cope with drought. While the exposure to water saving information fails to facilitate behavioral responses to address drought by an indirect path (via drought risk percep tion), it directly affects daily action responses to save water . Also, the familiarity with a drought policy has a positive impact on both daily action and device utilization responses . ound bit es and to educate the public about climate change ( Berk & Schulman, 1995, p. 31) . W hen facing drought, governments have to crystallize the expected intensity and duration of drought, detailed response plans (e.g., limited water supply and redistribution of 104 Dessai and Sims (2010) also regard effective communication as address drought. 105 Lorenzoni and Pidgeon (2006) responses to climate change. As mentioned before (see Footnote 57), this dissertation briefly accounts for the role of social norming in the next cha pter.

PAGE 139

125 water among domestic, industrial, and agricultural areas) by scenarios, water savi ng p ractice guidelines for citizens, and policies of water saving promotion (e.g., subsidies for households which use less water than usual) or water consumption deterrence (e.g., temporary water price hike and fines for households which use more water tha n usual). And t hen, along with public, private , and civic organizations , governments should conduct activities, such as education programs, campaigns, and p ublic discussion, to share those pr actical wisdom with the public. 106 Third, different socio demographic characteristics require different strategies. Age, education, income, and area are significant influential factors on daily action or device utilization responses related to water saving. T o maximize the effects of ral responses, governments must consider different r esponses derived from s ocio demographic features when designing educational programs, water saving promotion and water consumption deterrence po licies, and other activities. 107 Even thou gh this analysis does not deal with the interaction effects of in dependent (and socio demographic) variables, revealing these effects might help governments better understand individuals behavioral responses to address drought and prepare tailored strateg ies for individual groups with fine grained socio demographic characteristics. In sum, to cope with dro ught effectively perception oral responses to address drought in South Korea . The findings of this local level analysis make it clear that drought policy should be shifted from ready made (depending on risk perc eption) to tailor made approaches. 106 Piwowarczyk et al climate change. 107 Similarly, Messages about climate change need to be tailored to the needs and

PAGE 140

126 Limitations and Future Research Directions Some limitations need to be acknowledged. A primary limitation relates to a survey period. The K water survey in north west Chungnam was conducted between 10 and 23th December, 2015 , during the midst of the drought . Normally, a survey period of two weeks is not problematic in analyzing data . However, the threat of drought can become significantly aggravated day by day. This short time gap between initial and late respondents may influence the ir answers associated with drought risk perception and behavioral responses. 108 A second limitation is a lack of consistency in measuring variables. While some variables, such as behavioral responses and drought risk perception, are measured by several items, other s depend on a single item . The final caveat is the difficulty in extend ing the findings to other drought prone areas. This analysis focus es on north west Chungnam where the Boryeong dam is a main source of domestic water. The types of main water sources (e.g., dam, river, and grou nd water ) in areas may lead to different responses of individuals in addressing drought. Regarding future research directions, ceaseless e fforts for widening the research scope and elaborating survey instruments are not only valuable for the national level analysis but also this local level analysis. Additionally, m ore research projects have to actively utilize open ended questions and qualitative methods in analyzing this topic. N ot only revealing the causal relationship among personal factors, risk perception, and behavioral responses related climate change, but also Given that most previous studies related to this topic have depended on quantitative methods to verify causalities , taking a different angle (i.e., utilizing a qualitative approach) is likely to help not 108 There is a possibility that while the respondents in the early stage of the survey were hesitant to take action against drought even they considered drought as a serious problem, the respondents in the last stage of the survey were active in behavioral responses with high drought risk perception.

PAGE 141

127 only fill the gaps in existing quantitative research , but also find novel information. In doing that, i nterdisciplinary research, including scholars from diverse fields (e.g., sociology, economics, and psychology) as well as practitioners from public agencies, may contribute to achieving mor e meaningful results . 109 Conclusion The purpose of this analysis was to explore t he relationship among personal factors, drought risk perception, and behavioral respo nses to address drought at the local level . Data were deri ved from a face to face survey conducted in South Chungcheong Province , South Korea . SEM r egression analysis was conducted hypotheses and shows mixed results. Unexpectedly, drought risk perception do es not have a significant impact on either daily action or device utilization responses to save water . Even though thre e factors (i.e., awareness of water quantity condition, the expos ure to water saving information , and the assessment of water infrastructure policy ) are significantly correlated with drought risk perception , they also fail to have an indirect . Meanwhile, two knowledge factors (i.e., the exposure to water saving information and familiarity with drought policy) and four control factors (i.e., age , education, income, and area) directly affect daily action , device utilization , or both responses to save water . W hile self vulnerability perception associated with limited water supply shows a sig nificant relationship with device utilization responses, its direction is In light of this analysis , there is no clear evidence that drought risk perception behavioral responses to adapt to drought. 109 To obtain better ideas for addressing climate change, collabo rative research projects (between scholars and practitioners, between natural and social scientists, etc.) are recommended by many previous studies in the climate change field (e.g., Gifford, Kormos, & McIntyre, 2011; Le Dang et al ., 2014 ; Schmidt, Striegn itz, & Kuhn, 2014).

PAGE 142

128 However, keep in mind that different contexts (e.g., t he timing and period of a survey and the types of the main water sources in areas ) are likely to create different results. In other words , it is too early to underestimate the role of risk perception in addressing drought. Further research projects in diverse contexts and with much more refined instruments are needed to unlock the effect s havioral responses to cope with drought . Additionally , qualitative and interdisciplinary approach es will help seek missing parts of the findings which are mostly based on quantitative a nalyses . Regardless of whether drought risk perception works on behavioral responses, m ulti faceted communication (e.g., between governments and citizens and between citizens and citizens ) , the provision of more specific and practical water saving wisdom, and tailored strategies for different socio demographics might be recommendable methods for the enhancement of indiv responses to address drought . D espite some limitations, this study asserts that the findings of the analysis provide important insights that help practitioners design drought policy fitted in their aga inst drought .

PAGE 143

129 CHAPTER VI EXPERT LEVEL ANALYSIS T he findings of th e national and local level analyses not only help understand the causal relationship among personal factors, risk perception, and behavioral responses related to climate change, but also provide valuable policy implications. Meanwhile , it is difficult for the quantitative anal yses eses and the analysis results and to convey both the the meaningful stories that help pol icymakers refine climate policy in interpreting their findings. Thus , i n this last analysis, the research question is investigated more in depth based on expert knowledge . This dissertation argues that experts can not only provide insights on the causal relationship among the variables , but also insights on other factors affecting the relationship among them (e.g., cognitive factors or national contexts which hamper the relationship ) . 110 Interviews and survey data of climate change experts are analyzed from a qualitative approach question and help interpret the results of the quantitative analyses. Research Context s Necessity of Mixed Methods Research In response to a long and fierce debate between the supporters of quantitative and qualitative research paradigms, the necessity of mixe d met hods research incorporating both has been raised by many researchers (e.g., Almalki, 2016; Creswell et al., 2003 ; Johnson & Onwuegbuzie, 2004; Johnson, Onwuegbuzie, & Turner, 2007; Riccucci, 2010). 111 Mixed 110 Actually, experts are different from the general public. Nevertheless, climate change experts, who have scientific the gener al public, can help answer the research question and interpret the findings of the previous two chapters. 111 Riccucci (2010) describes the term paradigm Rainey (1994).

PAGE 144

130 Johnson & Onwuegbuzie, 2004, p. 21). 112 Similarly, Creswell (2009) goes on to combination of both qualitative and quantitative research than either form by itself. Their combined use provides an expanded underst . 113 It does not mean that all research projects can benefit from combining quantitative and qualitative methods ( Tashakkori & Creswell, 2007) . The choice of a research method should be determined by whether it is appropriate for answering research questions or not ( Carey, 1993; Creswell & Plano Clark, 2007; Johnson & Onwuegbuzie, 2004) . In other words , research questions reflecting both qualitative and quantitative aspects are subject to such mixed methods research ( Tashakkori & Creswell, 2007) . This dissertation the causal relationship among personal factors, risk perception, and behavioral resp onses associated with climate change. I t primarily focuses on the verification of the causal relationship among them. However, other issues, such as revealing why some rel ationship s are insignificant, identifying what stories are appropriate to explain unexpected results , and finding how those relationships can change over time , are also in tegral parts of solving the research question. Therefore, this study can answer the research question more successfully by mixing quantitative with qualitative approaches. 112 According to Johnson and Onwuegbuzie (2004), the core of ix[ing] or combin[ing] 17). Johnson, Onwuegbuzie, and Turner (2007) introduce diverse definitions of mixed methods research suggested by ma ny leading researchers. 113 However, mixed methods research also has disadvantages. It can be more time consuming, expensive, and difficult to carry out than a single method research ( Johnson & Onwuegbuzie, 2004) . Also, further investigation is needed when d ifferent results are detected within the data sets ( Almalki, 2016) .

PAGE 145

131 Advantages of Utilizing Experts In a qualitative approach, data are collected by several methods such as participant observation, documentary analysis, and in depth in terviews (Rubin & Rubin, 2012). According to Kvale (1996; 2003), qualitative interview s are a powerful method of obtaining quality information to address research questio Therefore, t his analysis garners data from in depth interviews with experts. 114 Compare d to participant observation and ordinary interviews, there is a high possibility that expert interview s can be conducted with less resources (e.g., time and money) and obtain more comprehensive and authoritative information from a relatively small number of interviews, respectively ( Bogner, Littig, & Menz, 2009; Libakova & Sertakova, 2015; Meuser & Nagel, 2 009) . In other words , expert interview s are the best m ethod to gather quality information swiftly, efficiently, and intensively. Expert knowledge, which Martin et al . (2012) research, such as defining problems, selecting variables, developing models, and interpreting results and characterizi ng uncertainty ( Martin et al ., 2012; McBride & Burgman, 2012). 115 More importantly, expert knowledge is a desirable method to resolve issues surrounding uncertainty and a lack of data ( Kuhnert, Martin, & Griffiths, 2010; Martin et al ., 2012; McBride & Burgman, 2012) . T here are also several risks related to expert interviews. For instance, like laypeople, 114 Interviews have been widely used by the humanities and technical studies with various topics ( Libakova & Sertakova, 2015) . Since the 1990s, expert interviews have been grown as one of qu alitative methods ( Meuser & Nagel, 2009) . 115 Krueger et al the preliminary state they also note that the distinction between expert knowledge and opinion is not clear.

PAGE 146

132 experts are subject to various biases, such as cognitive (e.g., anchoring and inconsistency) and motivational (e.g., group think and wishful thinking) biases ( Booker & McNamara, 2004; Krueger et al ., 2012 ; et al ., 2006 ) . 116 117 Also, t he findings of expert interviews can be in conflict with those of quantitative analysis based on empirical data ( Martin et al ., 2012) . Nevertheless, expert interviews have a potential to enrich the findings of this study, under the premise of selecting experts and conducting interviews rightly. Core Components of Experts well understood by people and frequently used in their daily lives. 118 M any research projects in diverse fields, such as politics ( Benoit & Laver, 2007; Ray, 1999) , technology ( Buchholz, Luzadis, & Volk, 2009; Catenacci et al ., 2013) , and environment (Krueger et al ., 2012; McBride & Burgman, 2012), have actively utilized experts to investigate their research topics. Martin et al . (2012) describe an expert as Krueger et al . (2012) de a condition for an expert (p. 4). Given these existing definitions, the answer to t he k nowledg e represents these factors) , but also on their relevance to the research topic . 116 According to Booker and McNamara (2004), anchoring refers to depending heavily on the first piece of information when solving a problem. Inconsistency can be triggere d by confusion related to assumptions or definitions, memory problems, and fatigue. Group think and wishful thinking are related to dysfunctional Booker and McNamara (2004, 117 To minimize these biases, several techniques, such as calibration and weighing, are recommended by previous studies (e.g., Ayton & Pascoe, 1995; Bo oker & McNamara, 2004; Krueger et al ., 2012). 118 In light of general definitions of an expert provided by many dictionaries (e.g., Collins dictionary (https://www.collinsdictionary.com/) and Oxford dictionary (https://www.oxforddictionaries.com/) ) , an exper t might be a person who has comprehensive knowledge, experiences, or skills in a particular field.

PAGE 147

133 Expert Interview s Data Collection S ubjects . Regarding climate change, the t ypology of experts can vary, such as meteoro l ogists, ecologists , hydrologists, and national and local public managers, depending on its specific sub topics. The experts of the sub topics linked to scientific and technological knowledge c an be identified with relative ease. However, those of the sub topics associated with a hybrid of scientific (or technol ogical) and societal knowledge cannot be easily detected. T he When selecting experts for interviews, this analysis beha vioral responses address ing climate change. The right choice of experts is a prerequisite for acquiring reliable and quality information. Informants should have both of the above mentioned knowledge. Furthermore, it is worth noting that the latter can be m ore powerful by accumulating many interperso n al experiences with the general public, rather than depending on academic learning. In this regard, this analysis pays attention to K water. 119 From a domestic perspective , K water operates numerous water related facilities, such as sixteen multi purpose dams, integrated regional water supply systems, and twenty two local water supply systems (K water, 2018) . Also, it acts not only as the number one renewable energy enterprise by providing about twenty five percent of the national facility capacity of renewable energy, but also as the leading company in developing eco friendly waterfront cities (K water, 2018). From an international perspective, K water has aggressively advanced its overseas 119 In 1967, K water was founded as the Korea Water Resources Development Corporation in accordance with the the heavy chemical industries and manufacturing (K wa ter, 2018). In 1974, it was reorganized into the Industrial Sites and Water Resources Development Corporation (K water, 2018). Finally, in 1988, K water was reformed into the Korea Water Resources Corporation Government Legislation, 2015b).

PAGE 148

134 business since 1994. It and has undertaken 32 overseas projects in 21 nations, including the construction of the Partrind Hydroelectric Dam in Pakistan (K water, 2018) . Overall, K water, founded in 1967, is th e only water specialized organization in South Korea as well as o ne of the leading organizations in global water management . Also, its business sectors are deeply associated with climate change. The K water institute, affiliated with K water, has many researchers from diverse fields, such as public policy, economics, business, ecology, meteorology, and civil, mechanical, electrical engineering. 120 Based on these professional personn el, it has conducted numerous research projects related to climate change, including extreme weather events (e.g., flood and drought). Among them, several studies deal with scientific and technical knowledge as well as the general publi c (e.g., survey and interviews). R esearchers who participated in those studies might have re sponses against climate change. Overall, this study views researchers in the K water institute who have been involved in climate change related studies as proper experts on the research topic. Given that they have different academic backgrounds, interviews with them can offer us chances to identify diverse perspectives derived from different fields concerning the relationship among personal factors, risk perception, and behavioral responses related to climate change. Data collection instruments . T his analysis garners data from in depth semi structured interviews with climate change experts. 121 Based on the interview protocol developed by the 120 This institute consists of six research centers: Research center for water policy and economy, water resources research center, infrastructure research center, water supply research center, wat er facility research center, and water quality research center (K water institute, 2018). 121 Campbell et al . (2013) view in depth semi structured data as (p. 295).

PAGE 149

135 researcher of this study , expert i ntervie ws are conducted by the researcher . Appendix I presents the interview protocol . 122 The identification of the interviewees is based on snowball sampling , which involves current interview subjects recruiting the next round of interview subjects ( Biernacki & Waldorf, 1981). The first interviewee is one of the researchers involved in the K A study on the strategic plan of water and energy sector for new which was conducted from 2016 to 2017. The number of the identified subjects are thirteen, and all of them have a PhD degree . 123 T hey were interviewed face to face in the conference room of the K water institute between 29th October and 3rd November, 2018 . The time required for each interview was about one and a half hour s. I nterview content was recorded by a voice recorder . Data A nalysis Coding . A ccording to Patton (2002), reducing the volume of raw information, sifting trivia from significance, identifying significant patterns, and constructing a framework for analysis which deals with massive amounts of data (p. 432). To effectively analyze d atasets derived from the interviews, the study utilizes coding which is viewed as a desirable method in ( Miles, Huberman, & Saldaña , 2014). 124 Considering the research variables , the researcher of this dissertation develops a codebook and continually refine s it during the analysis phase. Appe ndix J shows a major part of the codebook . 122 questions, appropriate follow up questions, which help draw detailed information from interviewees (Rubin & Rubin, 2012), are employed to successfull y achieve the above mentioned goals. 123 Eleven interviewees belong to the K water institute. The other two interviewees belong to a university and a national institute, respectively. 124 labels that assign symbolic meaning to the descriptive or inferential information compiled during Miles, Huberman, & Saldaña

PAGE 150

136 When cod ing data, this analysis employs a computer assisted qualitative analysis software, specifically NVivo software, to e nhance the quality of analysis. NVivo is regarded as a useful tool when conducting qualitative data analysis due to several advantages such as reducing manual tasks, facilitating organization of the data, and enhancing team worki ng (e. g., Azeem, Salfi, & Dogar, 2012; Buchanan & Jones, 2010; Hilal & Alabri, 2013 ; Wong, 2008). The coding Intercoder reliability . R eliability is important not only for quantitative research , but also for qualitative research (Golafshani, 2003). 125 R egarding the use of coding, many existing studies (e.g., Hruschka et al ., 2004; Tinsley & Weiss, 1975) highlight the importance of i ntercoder reliability. Simply, it refers to the extent to which different coders code the same data and arrive at the same conclusion (Lombard, Snyder Duch, & Bracken, 2002). It is assessed by Scott, 1 955). To test intercoder reliability, part of the expert in terviews (i.e., four interviews) was coded by a second coder , who has studied public affairs in a PhD program. measuring intercoder ag reement for categorical scales based on t he data of t he original coding and second coding indicate s that intercoder reliability is quite acceptable (more than .80) . 126 Analysis . The expert interviews include thirteen cases (individuals). To begin with, the importance of individuals role in addressing climate change and sig nificant personal factors affect ing risk perception (or behavioral responses) linked to climate change are identified by 125 Krippendorff (2004) introduc es three types of reliability: S tability, accuracy, and reproducibility. The latter is called intercoder reliability (see Krippendorff (2004) for more information). 126 Regarding an acceptable level of intercoder reliability, there are no coherent standards ( Campbell et al ., 2013; Lombard, Snyder Duch, & Bracken, 2002). Instead, based on the reviews of various rules of thumb su ggested by previous studies (e.g., Banerjee et al would be

PAGE 151

137 immersing in each case. Subsequently, the general relationship among personal factors, risk p ercept ion, and behavioral responses related to climate change, supported by the majority of the experts, is described based on the comparisons of all cases. Lastly, the iterative process of comparisons in all cases (moving back and fo rth between each and cross interviews ) is executed to discove r significant issues or stories that contribute to hypotheses and the results of quantitative analyses. By utilizing the se two complementary analytical processes (i.e., within each case and cross cases), this chapter can achieve more robust and reliable findings. 127 Analysis R esults The results of the qualitative analysi s consist of three parts. F irst, the importance of is evaluated . Second for each hypothesis is explained. The final section of the results will discuss how the findings from the interviews can help fill the gap s between the hypotheses and the results of the quantitative analyses at the nati onal and local levels . Importance of i . As mentioned earlier, many previous studies (e.g., Bostrom et al ., 1994; Leiserowitz, 2006; Ruddell et al ., 2012) highlight that individuals can play a significant role in coping with climate change by perceiving its risk and taking action. The results of the expert interviews also correspond to their view. All interviewees (hereafter to be referred as EI, from change is either very or somewhat important. Some crucial points related to their role eme rge . First, some experts recognize that 127 Ayres, Kavanaugh, and Knafl (2003) introduce within and across case analytic strategies for qualitative studies.

PAGE 152

138 two experts acknowledge that policymakers in South Korea justify the allocation of li mited resour perception s and responses to climate change Second, six of the interviewees believe that individual behaviors are essential for both mitigating and adapting to climate change . Without se curing individual (e.g., following policies, purchasing goods provided by GHG reduction oriented companies, and utilizing ite ms which minimize GHG emissions) , the efforts of governments and ind ustries against climate change would not be effective. One interviewee states, add up to a great solution to Third, seven experts feel that despite the importance of still playing the leading role in a ddressing clim ate change . Instead, all interviewees recognize that designing and implementing cli mate policies. They believe that t hose behavioral responses help governments not only alleviate climate change without enforcing strong regulations to the general public and industries, but also save valuable resources (e.g., time and money). Last, two interviewees predict that the role of individuals in combating climate change Additionally , some others discreetly point out that benefits derived from the role of individuals can only be reaped from the proper mitigation and adaptation guidelines ]. Also, two experts warn that remarkable impacts on the mitigation and adaptation of climate change can only be realized not by behavioral responses of few individuals, but by behavioral responses ].

PAGE 153

139 Together, the role of governments is considered more important than that of individuals in addressing climate change so far. Nevertheless, it is obvious that individuals play a significant role in fighting climate change by influencing both governments and industries, and their behavioral responses have a potential to release the burden of governments. More importantly, mass participation, its impact can be strengthened and contribute more to achieving effective and se who are directly affected by the adverse impacts of climate change. Therefore, their more proactive and voluntary activities to s one of the experts Expert perspectives on the hypotheses . To obtain reliable and specific answers, the expert interviews considered both the mitigation and a daptation contexts. The experts provide better understand some of the more nuance results of the national and local level surveys. Table 9 In terms of the mitigation context, most interviewees recognize that personal factors significantly affect risk perception and higher risk perception , along with greater poli cy related knowledge, intensifies behavioral responses . The core hypothesis, the positive relationship between risk perception and behavioral responses, is supported by nine interviewees. However , some others believe that several barriers , such as financial cos ts and inconvenience in daily life , interrupt the relationship between risk perception and behavioral responses. Regarding other hypotheses, th irteen, eleven , and nine interviewees advocate that risk perception is influenced by short term expe rience s , general knowledge , and policy assessment , respectively. U nexpectedly, the majority of the experts do not regard long term experience s as an influen tial factor on risk perception.

PAGE 154

140

PAGE 155

141 In the case of the adaptation context, it is doubtful whether drought risk perception is a sponses against drought. T he majority of the interviewees agree with three hypothesized relationships between personal factors (i.e., water problem experience, ge neral knowledge, and policy assessment) and risk perception. However, the core hypothesis that higher risk perception promotes behavioral responses is supported by only five interviewees. The experts opposing this hypothesis argue that certain conditions, such as countermeasures agai nst prolonged drought ( e.g., multi purpose dams and multi regional water supply systems), and inconvenience in changing behavioral pattern s during drought, undermine the influence of risk perception on behavioral responses to drought. In addition, the positive relationship between policy related knowledge and behavioral responses is identified by nine interviewees both in the mitigation and adaptation contexts. Qualitative insights for each hypothesized relationship are described in the next sub section. Overall, from a dichotomous approach whether the majority of the interviewees support each hypothesis or not, personal factors, except long term exper i ence, can significantly affect risk perception (or behavioral responses) both in the mitigation and adaptation contexts. However, while risk perception can influence behavioral responses in the mitigation context, the role of risk perception as a mediator is quite uncertain in the adaptation context. It is clear that these findings, along with the findings of the expert survey, help accumulate knowledge about the causal relationship among personal factor, risk perception, and behavioral responses related to climate change. Nevertheless, those are still insufficient to interpret the re sults of the national and local level analyses. More detailed information derived from the in depth interviews is subsequently presented to complement the results of two quantit ative analyses.

PAGE 156

142 Filling the gap s between the hypotheses and the results of quantitative analyses . Expectations and results do not always go together. The results of two quantitative analyses presented in the previous chapters also do not perfectly correspond to the hypotheses of this dissertat ion. Ironically, the discrepancies between the analysis results based on statistical metho to a more careful approach to the research question and makes the findings of the e xpert interviews all the more worthwhile. This sub section, the heart of this chapter, tries to fill the gap s between them. Short term climate change and water problem experiences . hypothesis, both the national level and local leve l analyses, based on a quantitative approach, fail to identify the positive relationship between experiences and risk perception. Short term experiences and water problem experiences do not statistically influence climate change risk perception and drought risk perception, respectively. They also do not affect behavioral responses to address climate change. However, it would be rash to conclude that experiences are not an influential factor on risk perception before unveiling the reasons for this discrepanc y. Fortunately, some clues to explain the discrepancy are reflected in the expert interviews. term experiences positively affect their climate change risk perception. Likewise, eleven interviewees support the positive relationship between water problem experiences and drought risk perception. Definit ely, experiences are regarded as one of the strong influential factors on risk perception by climate change experts. But, many of them stress that those relationships can happen under certain conditions. For examples, the majority of experts believe that i ndividuals must experienc e extreme weather events with dire ct and si gnificant damage before they are likely to respond to climate change . Three interviewees recognize that individuals must feel uncomfortable for quite a long time due

PAGE 157

143 to water problems . relationship between experiences and risk perception is significant depends on the extent of . Also, six of th e experts argue that individuals must perceive extreme weather events as part of clima te change . Additionally , assuming that short term and water problem experiences do not statistically influence climate change and drought risk perception in empirical ana lyses respectively, interviewees wer e asked to explain why. Table 10 lays out the possible reasons for the Table 10 Possible reasons for the insignificant relationship between experiences and risk perception National level analysis Local level analysis Possible reasons Respondents Possible reasons Respondents They might experience extreme weather events with minor damage (or damage within manageable levels) seven others] Their discomfort brought by water problem experiences might not be significa nt and may not last a long time They might not know that extreme weather events they experienced are climate change five others] Water problem experiences without proper information about drought might increase not drought risk perception, but discontent with government As time passes, higher climate change risk perception triggered by extreme weather event experiences might return to baseline As time passes, higher drought risk perception caused by water problem experiences might be back to its original state After extreme weather event experiences with severe damage, proper countermeasures might be implemented by governments There might be countermeasures against drought, such as multi purpose dams, multi regional water supply systems, etc. For individuals who too frequently experience extreme weather events, those events might be taken for granted and might not be regarded as climate change Water problem experiences can be triggered not only by drought, but al so by other events (e.g., pipeline replacement and facility faults)

PAGE 158

144 All told, while short term and water problem experiences have the potential to positively affect climate change and drought risk perception respectively, their potential is likely to be actualized under certain conditions, such as experiences accompanied by direct and significant damage (discomfort), experiences with appropriate knowledge about climate change (drought), recent experiences, and instances with no available solution. Given the responses of the in terviewees presented in Table 9 and 10 , it is possible that short term and water problem experiences positively affect climate change and drought risk perception in some countries respectively, it is not possible in other countries. In South Korea, individuals are generally exposed to information about climate change and drought derived from diverse sources, such as off line and on climate change spe cialized websites (e.g., Korea Information Portal for Climate Change A daptation http://clim ateadapt.kei.re.kr/portal/, M y w ater https://www.water.or.k r/, a nd National Drought Information Analysis C enter http://drought.kwater.or.kr/main .do), and social quality internet envir onment. Mea nwhile, some experts (e.g., ) point o ut that South Korea has a high level of water management capability, based on multi purpose dams and multi regional w ater supply systems , and therefore most individuals do not suffer serious dama term limited water supply (e.g., dozens of hours), but experience short term (e.g., a couple of hours) or night time limited water sup pl y. As shown in Figure 12 , in fact, the amount of damage caused by typhoons and heavy rains in South Korea has been on a downward trend since 2002 . 128 R ecently, there was no 128 Typhoon Lusa in 2002 triggered the greatest amount of damage among the typhoons, which occurred during 1987 and 2017, in South Korea (Ministry of the Interior and Safety, 2017).

PAGE 159

145 catastrophic damage triggered by them (Ministry o f the Interior and Safety , 2017). 129 In other words, while some survey respondents answered that they had short term or water problem experiences with damage, that damage might not be significant enough to intensify their climate change or drought risk perception. Also, the impact of their e xperiences on risk perception is less likely to be permanent. Note : The data are provided by Ministry of the Interior and Safety (2017). The exchange rate is USD 1 = KRW 1,000 Figure 12 . Amount of damage caused by typhoons and heavy rains in South Korea (From 2002 to 2015) Based on the findings of both quantitative and qualitative analyses, t his dissertation tries to describe changes in the relationship between short term experiences and risk perception rel ated to climate change in Figure 13 . Those changes are deeply associated with the levels of the f ormation networks . S pecifically , while countries with weak adaptation capacity and strong information networks are likely to show the 129 The national level and local level analyses employ the data of surveys conducted in the middle of 2016 and the end of 2015, r espectively. Therefore, Figure 12 presents the amount of damage triggered by typhoons and heavy rains from 2002 to 2015.

PAGE 160

146 strong positive relationship between them, countries with strong adaptation capacity and weak information networks may present the insignificant relationship between them. Both countries with strong adaptation capacity and information networks and countries with weak adaptation capacity and information networks are likely to show the insi gnificant relationship or weak positive relationship between short term experiences and risk perception associated with climate change. However, these arguments need much more empirical evidence from future research. Figure 13 . Changes in the relationship between short term experiences and risk perception related to climate change In light of Table 10 and the above mentioned South Korea n context (e.g., strong adaptation capacity and information networks), the statistically insignificant relationship between short term (water problem) experiences and climate change (drought) risk perception in the previous empirical chapters might be unde rstandable. Unfortunately , contradictory results of the

PAGE 161

147 relationship between drought experience and each behavioral response to mitigate climate change in the sub models of the national level analysis remain unsolved. 130 Long term climate change experiences . term climate change experiences do not significantly influence their climate change risk perception. Five interviewees recognize that while it is possible for individuals to experience long term climate change, they are likely to adapt to those incremental changes happening over a long time rather than perceiving higher cli mate change risk . Some of them xample. 131 This syndrome is often used as a metaphor for the failure of perceiving creeping risks. Also, one expert believes that with long However, the fi ndings of the national level analysis and expert survey reveal that long term climate change experiences significantly and positively make sense, this study asser ts that the hypothesis, identified by the results of the quantitative analysis and expert survey, are more convincing and reliable rather than the opposing opinion supported by those of the expert interviews. On the sidelines, while it is challenging to un veil what exactly makes those gaps, this effort might be meaningful to future research related to this topic. This study tries to explain the gaps by raising three plausible possibilities. First, individuals are likely to be more understanding 130 Some responses (e.g., utilizing renewable energy and donations to environmental organizations) are positively affected by drought exp erience. But, some others (e.g., saving energy and saving water) are negatively influenced (see p. 89). 131 Boyatzis (2007) describes this syndrome. jump out of the pot. But if the frog is placed in a pot of cool water, which is gradually brought to a boil, then the frog (p. 162).

PAGE 162

148 of the adver se impacts of climate change in the future caused by recent incremental changes they have experienced rather than what the majority of the interviewees think. Boiling frog syndrome, assuming a lack of understanding of negative consequences in the future, i s an extreme and inappropriate example in describing the relationship between long term experiences and risk term climate change experiences might have already reached (or gone beyond) a critical point. Some experts (e.g., assumption that long term climate change experiences are still below a critical point. When reaching the critical point, they will significantly intensify climate change risk perception. The definitions of a critical point may vary in different fields (e.g., Thermodynamics and Mathematics). This dissertation views that if most individuals believe that they can no longer easily ad apt to climate change and the negative consequences of climate change are more than possibility based on their long term experiences, their long term climate change experiences reach the critical point. Third, long term climate change experiences and clima te change risk perception affected by these experiences are less likely to travel the same route. One of the interviewees states that while climate change is incremental over a long period (e.g., a straight and gently upward sloping line), climate change r isk perception influenced by long term climate change experiences repeat the prolonged stability and radical change (e.g., horizontal lines with steep upward sloping lines) . This argument looks similar to punctuated equilibrium ry of policy dynamics as it focuses on the mechanisms that lead to policy Jones & Baumgartner, 2012, p. 4) and explains both stability and radical change in the policy process ( Baumgartner, Jones, & Mortensen, 2014). Regrettably, the expert interv iew does not provide detailed information about the mechanism, such as what factor triggers the radical

PAGE 163

149 change and when it occurs. While the majority of the interviewees only consider the stable stage of climate change risk perception, the change stage of that also exists. Both the significant or insignificant relationship between two variables may be possible depending on which stage climate change risk perception affected by long term climate change experiences is going through. Overall, all of these are mere possibilities, lacking theoretical and empirical foundations. Further research to verify those possibilities can help make a research model for elaborate. Genera l knowledge . Not only can knowledge associated with a risk help individuals notice the existence of the risk, but it can also help them specifically predict the negative effects of the risk says one expert Most interviewees agree that climate change and drought knowledge positively influence climate change and drought risk perception, respectively. Additionally, some of them point out several cautionary notes in understanding the relationship between risk re lated knowle dge and risk perception. First , without direct or indirect experiences, higher risk perception triggered by knowledge is likely to decrease gradually as time goes by , sometimes knowledge intensifies dissatisfaction with the gover there is a possibility that too much scientific knowledge leads individuals to make a mistake of underestimating the adverse impacts of , it is necessary to check whether individuals have The findings of the expert interviews are consistent with those of the two quantit ative analyses with respect to the relationship between climate change knowledge, risk perc eption, and behavioral responses. Given these results , the interviewees were asked what type of knowledge

PAGE 164

150 that could intensify public perceptions of climate ch ange and drought risk . Ten of the interviewees recognize that short term and scientific knowledge in raising those risk perceptions. Some examples of t he short term and practical knowledge suggested by the experts are as follows: Presenting expected damage rather than global and national d amage if Spotlighting significant but hard to notice changes caused by climate change in the past Sharing both success and failure stories related to climate change adaptatio n rather than adaptation (Perfect adaptation is impossible) [ , Providing guideline for al responses to address climate change (Not only can this knowledge help take action to address climate change, but it can also help Overall, climate change and drought knowledge are significant and reliable factors in intensifying climate change and drought risk perception, respectively. Further, when sharing short term and practical knowledge rather than long term and scientific knowledge, individuals might perceive the risk of climate change more clearly and vividly. To strengthen the impacts of

PAGE 165

151 Policy assessment . This factor, along with short term experience and knowledge, is regarded by the majority of the interviewees as a significantly influential factor on climate change and drought risk perception. Most experts state that the levels of policy assessment are negatively associated with those risk perceptions. But, they are also wary of some latent obstacles distorting the relationship between them. For example, there is a possibility that he levels of the policies . There is also a chance that individuals assess climate or water policies inattentively without careful consideration [EI 1]. The results of the expert interviews are mixed results. While the assessments of drought policy and water infrastructure policy significantly and negatively affect climat e change and drought risk perception respectively, those of flood policy and water supply policy do not. Regarding such minor discrepancies between the hypothesis and the results of empirical analyses, some interviewees note that assessing the response to damage from an extreme weather event is more likely to show a significant relationship with climate change risk perception when the extreme weather event has the following characteristics: Getting serious recently and being exposed to the med Having difficulties in creating response infrastructures [EI 8 and EI 12] In South Korea, th e amount of damage caused by typhoons and heavy rains has been dec lining for a while (see Figure 12 ). Also, property damage by floods can be indemnified by

PAGE 166

152 taking out storm and flood insurance ( Ministry of Government Legislation , 2017 a ) . 132 Meanwhile, droughts have occurred more frequently in the 2010s ( National Drought In formation Analysis Center , 2018), and they have attracted more attention from the media. 133 Droughts typically affect a wider area and continue longer than floods, and a high level of wate r supply infrastructure is necessary in effectively addressing prolonged droughts. In light of these things, the mixed results of the quantita tive analyses are understandable . Overall, while ct climate change and drought risk perception, the influence of each assessment on those risk perceptions may vary with the above mentioned features, and sometimes that influence even disappears. Policy related knowledge . familiarity and their behavioral responses to cope with climate change (drought) is supported by behavioral r havioral responses finally benefit themselves based on higher levels of policy familiarity familiarity with policies combined with incentives is likely to be more effective in enhancing ngs of the quantitative analyses and expert 132 from the state or a local government (Ministry of Government Legislation, 2017a). 133 BIGKinds ( https://www.bigkinds.or.kr/) , the search engine for news, operat ed by the Korea Press Foundation (http://www.kpf.or.kr/) helps analyze news from fifty three media outlets, including broadcasting companies and metropolitan newspapers, in South Korea. When searching BIGKinds for drought under detailed conditions (i.e., k eyword drought, topic category society and local, and accident category drought among natural disasters), the number of the searched news is increasing sharply from 767 in 2010 to 4,707 in 2015.

PAGE 167

153 discrepancies are observed among them. Risk perception . In terms of climate change risk perception, some interviewees argue t hat hange risk perception may not always result in behavioral responses related to climat e change , due to several obstacles , interviewees are convinced that climate change risk perception intensifies behavioral responses against climate change. 134 The results of the quantitative analysis in the national level chapter also identify the significant relationship risk perception and their behavioral responses. 135 In this regard, climate change risk perception affected by personal factors plays a pivotal role as a mediator in the nses to mitigate climate change. However , in the case of drought risk perception, perception is positively associated with their behavioral responses to address drought is inconclusive and debatable. The e xperts were divided on this question between t wo groups ( i.e., the first group supporting the significantly positive relationship between them; the second group arguing the insignifi cant relationship between them) . Further, the results of the quantitative analysis at the local level fail to show a significant relationship between drought risk perception and behavioral responses (both daily action and device u tilization responses) against drought. To fill the gap results of the empirical analysi s, it is worth noting arguments of the experts who belong to the second group. These are as follows: 134 However, the time needed to bring out behavioral responses from risk perception may differ from individual to 135 This statement is based on the relationship between climate change risk perception and the number of behavioral responses to mitigate climate change.

PAGE 168

154 In severe drought periods, governments generally prioritize the supply of residential water over that of water for other purposes, and individuals still can use water wi thout any restraints, or with There are reliable countermeasures against prolonged droughts, such as multi purpose dams, multi Drought risk is suddenly Each individual generally thinks that his or her behavioral responses do not help It is difficult for individuals to suddenly change their patterns of water usage during every droug All of these arguments , which are largely consistent with Wachinger et al . (2013) and Bube ck, Botzen, and Aerts (2012), have a potential to neutralize the influence of drought risk perception on behavioral responses to address drough t. Especially, the first two arguments, mostly in line with those mentioned in the experience part, are viewed as the main culprits of making the relationship between them sta tistically insignificant in South Korea . Namely , it is difficult for drought risk perception influenced by personal factors to become a significant l responses against drought in South Korea. Still, drought risk perception may be an important factor in other cou ntries with different cont exts. T hese results imply that it is important to develop more specific hypotheses and elaborate theories in exploring the relationship between risk perception and behavioral responses related to climate change. In doing that, reflecting the different features of climate risks (e.g., long term climate change, drought, and flood) and different contexts within which these risks arise could be a good strategy.

PAGE 169

155 Self vulnerability perception related to limited water supply . W hile it has a statistically significant relationship with one of the behavioral response types (i.e., device utilization response) in the local displays a negative directi on. According to the interview s, it is al most natural for individuals who perceive higher self vulnerability to limited water suppl y to be reluctant to take water saving actions during drought period s from three points of view. Some expert s believe that t his may be because individuals perceiving higher self vulner ability use much water by nature and are not accustomed to water saving behaviors . Four interviewees recognize that they may also try to secure more water than they generally need before limited water supply. It may be a self defense mech hile limited water supply makes them very uncomfortable, it is uncertain whether this regulation will be implemented in the future. Meanwhile, taking water saving actions immediately inconveniences them [E Given these thing s, self vulnerability perception is likely to work in the opposite way from how climate change and drought risk perception work. Thus, exploring how to reduce this perception seems necessary. Alternative factor in adaptation : Social norms . S argue that the clue to the savin g actions during drought comes from keeping up with the Joneses. Existing studies, mostl y related to social norms, note that ( Graziano & Gillingham, 2014; Terry & Hogg, 2001) , by being informed of how most of their neighbors act in a given situation ( Allcott, 2011; Schultz, 1999) , or by beliefs of how actively their neighbors are engaged in certain behaviors ( Nolan et al ., 2008) . 136 136 Numerous research projects have investigated the impact of social norms (or normative social influence) on human behavior ( Allcott, 2011; Nolan et al sets of beliefs about the behavior of Schultz, 1999, p. 26). Cialdini, Reno, and Kallgren (1990) introduce two types of social norms: Descriptive norms

PAGE 170

156 Namely, there is a possibility that individuals a re mor e aggressively engaged in water saving actions during drought when they positively assess (or have strong beliefs in) the beh avioral responses of n eighbors. To empirically verify that possibility, this study adjusts the research model in the local level analysis by adding the following variable. In the local level survey, respondents were asked to evaluate the l saving behaviors during the drought period compared to non drought period on a five point scale, where 1 = strongly saving , 2 = moderately saving , 3 = no difference , 4 = moderately using more , and 5 = strongly using more . The scale o f the question is reversed. H igher values indicate greater water saving behaviors have a significa nt and positive relationship with both the ir daily action responses (b = .571, p<.05) and device utilizati on responses (b = .267, p<.001) . Overall, while climate change risk perception is considered by most interviewees as a significant influential factor on behavioral responses to cope with climate change, drought risk perception is not considered influential risk perception on behavioral responses to address drought becomes insignificant seem valid i n the South Korea n context. These results are largely consistent with the findings of two quantitative analyses. Namely , risk perception is a credible factor of intensifying behavioral responses in the mitigation context, but it is less likely to work well in the adaptation context. Instead of drought risk perception, water saving behaviors of individuals are statistically affected by those of their neighbors. Unfortunately , t he findings of sub models in the national level analysis are not treated in the expert interviewees.

PAGE 171

157 Expert Survey Data Collection S ubjects . While the above mentioned interviewees are regarded as appropriate experts on this study enhance the reliability of this qualitative analysis, a separate expert survey is conducted to members of the Korea Water Resources Association (KWRA) affiliated with diverse organizations. 137 n par with K knowledge on water related issues including flood and drought. Recall, the relationship between risk perception and behavioral responses related to d results of this survey can enrich the findings of other analyses. Data collection instruments . A self administered and Web based survey to members of the KWRA was conducted www.open on May 30, 2018. The survey questionnaire focusing on the mitigation context was developed jointly by the researcher of this study and researchers at the research center for water policy and economy in the K water institute. It is composed of eleven close ended and four open ended questions . Socio demographic Characteristics This survey using token financial incentives has one hundred twenty six respondents. Among the nine ty eight respondents, 69 have a PhD degree ( 70.4%), 17 are in a PhD program (17.4%), and 12 ssional career (12.2%). These respondents are regarded as appropriate experts and are subject to this analysis. They 137 The Korea Water Resources Association (http://www.kwra.or.kr/) aims to develop water resources research, planning, development, management, and education in South Korea. It has more than two thousand members who belong to different organizations, such as universities, private companies, and public agencies (KWRA, 2018).

PAGE 172

158 range in age from 24 to 83 years (M = 43.73, SD = 10.72 ) and 88.78% are males. 138 Forty eight respondents are practitioners (i.e., 13 from governments; 13 from public corporations; 22 from private companies), and fifty respondents are scholars (i.e., 11 from public or private research institutes; 31 from univers ities; 8 from research institutes affiliated with universities). On Analysis Results close ended questions. Regarding the impacts of climate change and the role of individuals in addressing climate change, more than 90% of the respondents answer that the adverse impacts of climate change are ( moderately or very te change is ( moderately or very ) important. However, around 66% of the respondents believe that the general public does not perceive climate change risk accurately in South Korea. In terms of the relationship among personal factors, climate change risk pe rception, and behavioral responses to experiences, knowledge, and low satisfaction with climate policy intensify their climate change risk perception. Similarly, the majority of the respondents (above 70%) agree that climate change risk perception and policy related knowledge positively influence behavioral responses both to mitigate and adapt to climate change. Lastly, most respondents either agree (64.3%) or strongl y agree (21.4%) that Table 11 summarizes the ended question. 138 It is almost impossible to identify the male female climate change expert ratio in South Korea. Instead, according to Park (2018), the percentages of male professors at universities in South Korea in 2018 are 74.1%. This expert survey seems to somewhat oversample males compared to the population.

PAGE 173

159

PAGE 174

160 As mentioned above, this survey includes several open ended sub questions. When the personal experiences ( H1 ) and behavioral responses triggered by risk percepti on ( H5 ), they were requested to write a brief description of their selections. In terms of the question 4 and 5, less intensify (or weakly intensify) their climate cha nge risk perception. It does not mean that they deny the relationship between climate change experiences and risk perception. Instead, they present certain conditions needed for the significant relationship between them. The important descrip tions are summ arized in Table 12 . In short, repetitive climate change experiences accompanied by direct and severe damages can significantly intensify climate change risk perception. However, many of them assert that extreme weather event experiences of most individuals are not repetitive and are linked to indirect and minor damages in South Korea. Table 12 . Descriptions about the relationship between climate change experiences and climate change risk perception Description Summary Respondents Climate change experiences accompanied by a direct damage can significantly affect climate change risk perception Climate change experiences with a minor damage do not significantly influence climate change risk perception Extreme weather event experiences from a long time ago do not significantly affect climate change risk perception Repetitive extreme weather event experiences can significantly influence climate change risk perception People tend to forget the seriousness of extreme weather events once those events are over. In the cases of the question 8 and 9, less than 10% of the respondents either disagree or strongly disagree behavioral responses to address climate change, due to several reasons such as social loafing ],

PAGE 175

161 and the discrepancy between higher climate change risk perception and negligible discomforts in 139 Overall, t he results of the close ended questions support ing are no t entirely consistent with the findings of the quantitative analyses . 140 Fortunately, t he results of of the open ended questions , along with those of the expert interviews, give clues as to why personal experiences did not clearly have a causal relationship wit h climate change risk perception and behavioral responses to address climate change in both the national and local level analyses and why drought risk perception did not significantly influence behavioral respons es against drought . The findings of the expe rt interviews and survey are wrapped up in the concluding section of this chapter, following the policy and research implications. Lastly , e ven though this dissertation tried to identify appropriate experts for the interviews and survey to examine the research question , there is still a possibility that some experts could misunderstand how the general public actually b ehave and could also engage in the above mentioned biases (e.g., group think and wishful thinking). 139 Social loafing Latané , Williams, and Harkins (1979, p. 822), means that individuals exert less effort when doing something in groups compared to doing that alone. 140 This dissertation conducts an ordered logit regression analysis to investigate whether the demographic characteristics influence their opinions. O rdered logit models are appropriate when data are discrete values in sequential order, but the exact distance between the response options is unknown ( Williams, 2016) . Dependent variable s are the results of the eleven closed ended questions. Independent variables are gender (0 = male ; 1 = female ), age ( years ), career ( years ), education (0 = ; 1 = PhD program ; 2 = PhD degree ), and a field of activity (0 = the academic world ; 1 = ). There is no statistically significant relationship between the dependent and independent variables, except only the question 6 and 7. In the case of the question 6 related to knowledge, for a one level increase in education, the odds of strong intensification versus moderate intensification (moderate intensification versus minor intensification or minor intensification versus no intensification) are 0.377 times lowers, given the other variables are held constant in the model (p<.05). In terms of the question 7 associated with policy assessment, for the odds of strong intensification versus moderate intensification are 3.854 times greater than for the experts in the academic world (p<.05 risk perception) affect the answers to the question 8 (i.e., the relationship between risk perception and behavioral responses to mitigate climat e change) and 9 (i.e., the relationship between risk perception and behavioral responses to adapt to climate change) is explored based on the regression analysis. Consequently, only the opinion on the is positively associated with the opinions on the question 8 (p<.01) and 9 (p<.01).

PAGE 176

162 Policy Implications The experts end their interviews by suggesting several viable methods to consolidate the require stepping up education, financial incentives and penalties, information, commu nication, infrastructure, and publicity. Not surprisingly, the most suggested method is the enhancement of education [ Eight interviewees ]. The interviewees argue that knowledge of climate change should be spread through the regular school curriculum, especially primary and secondary school curriculums, emphasizing the timing of education. This suggestion echoes one of the policy implications, raising the necessity of early and systematic education, in the national level analysis. The second most sugges ted method is the implementation of policies with financial incentives [ Six interviewees ], followed by that of policies with financial penalties [ Four interviewees ]. While both financial methods have a greater potential to change individual behavior s relat ed to climate change , which method is b etter for enhancing their behavioral res ponses is quite debatable. This study believes that it depends on the features of the policy . The provision of information is also regarded by some experts as a helpful method t o strengthen that securing high method effective. In addition, the facilita tion of communicating among stakeholders [EI 11 and d by a minority of the experts. It is true that n ot only th e first method , but also the others , mostly overlap with the policy implications sugg ested in the national and local level analyses, and they are not new and innovative. However, at the other end of the spectrum, it can be interpreted that puttin g greater

PAGE 177

163 efforts into those ordinary but academically and practically proven methods is a better idea than behavioral responses again st climate change. This study asser behavioral responses to address climate change mediated by risk perceptio n can benefit from those ways. Research Implications The expert interviews also provide some valuable implications for future research rela ted to this topic. First, it is possible that personal factors have a non linear relationship (e.g., an inverted U shaped curve) with risk perception or behavioral responses, depending upon their levels. For example, while knowledge is likely to have a pos itive relationship with risk perception until a certain level, it may have a negative relationship with risk perception above that level. Second, there is a chance that the impacts of personal factors on risk perception and behavioral responses change over time. For instance, a recent extreme weather event experience (or recent knowledge) may affect risk perception more significantly rather than old one. Third, individuals may show a different time lag between the formation of risk perception and the practi ce of behavioral responses. For example, while some immediately take action after perceiving climate change risk, others need some or quite a long time to change their behavior patterns for responding to climate change. Fourth, hypotheses about the relatio nship among personal factors, risk perception, and behavioral responses against climate change should be adapt to climate change. Unfortunately, these things canno t be revealed by this dissertation due to the limitation of data and are left to future research projects based on a more standardized and refined survey.

PAGE 178

164 Conclusion By conducting face to face interviews (2018, N=13) and an online survey ( 2018, N=98) with climate change experts from a qualitative approach , this chapter tries not only to obtain n s a nd the results of quantitative analyses. Looking at the big picture, the experts both in the interviews and survey behavioral responses to climate change. However, the discre pancies between the hypotheses and the results of quantitative ana lyses in the national and local level chapters revealed some limitations in the literature guiding these hypotheses. Fortunately, the experts offer clues to these challeng es. Three considerable discrepancies and expert explanations for these phenomena are as follows. First, short term and water problem experiences show an insignificant relationship with climate change and drought risk perception, respectively. While most ex perts both in the interviews and survey advocate that short term experiences can intensify climate change risk perception, they commonly voice concern that this relationship is likely to disappear under certain conditions such as experiences with indirect or minor damage, experiences with a lack of or inappropriate knowledge, old experiences, and instances with a viable solution. Similarly, the relationship between water problem experiences and drought risk perception may be affected by these conditions. Se cond, the relationship between the assessment of policy and risk percept ion leaves mixed results in both quantitative analyses. The negative relationship between policy assessment and risk perception is also supported by mos t experts both in the interviews and survey . However, according to some interviewees , if an extreme weather event is serious and

PAGE 179

165 receives significant media attention, damages a wide area for a long time, and these damages are not compensated , the assessment of policy will s ignificantly influence climate change risk perception, and vice versa. Also, some experts believe that the core concern of the general public experiencing prolonged drought is not the level of water supply service, but that of water supply infrastructure. The latter is more likely to have a significant relationship with drought risk perception. Lastly, drought risk perc eption fails to intensify water saving actions. Unlike climate change risk perception, the role of drought ris k perception in promoting wate r the expert interviews disagree that drought risk perception can intensify behavioral responses to address droughts. They highlight that individuals generally experience minor damage from droughts due to prioritized residential water supply policy and established response infrastructure. Also, it is not easy for individuals to change their patterns of water usage whenever drought is happening, and an unexpected rain often relieves drought abruptly. As for the South Korea n context, the amount of damage brought by typhoons and heavy rains has decr eased over the last decade (Ministry of the Interior and Safety , 2017). A higher level of water management has been achieved through multi purpose dams and multi regional have been well established. The government has prioritized residential water supply during drought periods. B ased on the strong adaptation capacity , information networks, and severe damage from extreme weather events a nd very long term limited water supply triggered by drought. But, it is also clear that droughts have been getting serious and frequent and have

PAGE 180

166 draw n attention from the media (National Drought Information Analysis Center , 2018). In light of these facts, t he above mentioned discrepancies may come as no surprise. Overall, given the r esults of the expert interviews and survey , this study finds support for the expectation that personal factors affect risk perception an d higher risk perception will lead to behavioral responses to climate change. However, what is important to know is that this process does not operate perfectly and is influenced by the context of each country (or region or locality ) . Additionally, the expert interviews offered several sug gestions for policymakers, such as the enhancement of education, financial benefi ts, and publicity responses to cope with climate change . Some research implications, such as non linear relationships among variables and time eff ects on those relationships, were also presented to encourage further studies related to this topic. One of the limitation s of this research is whether the survey and interview participants adequately represent all experts relate d to this topic in South Korea, or whether the experts fully understand individual level risk perceptions and behaviors. Those limitations might weaken the validity of this analysis .

PAGE 181

167 CHAPTER VII CONCLUSION Although past decades, it is still getting worse due to insufficient global efforts in reducing GHG emissions . According to previous studies ( e.g., Clayton et al ., 2015; Semenza et al ., 2008; Shove, 2010), i can be alternatives for alleviating the negative impacts of climate change . Further , voluntary behavioral responses to address climate change can ease the financial and legal burden of policymakers in designing and implementing climate policies , including rules and regulations . So, unveiling the limate change is a very meaningful challenge both academically and practically. Personal factors and c l imate change risk perception have drawn much attention from scholars in that they have the potential to intens address climate change. However, empirical evidence about the causal relationship among personal factors, risk perception, and behavioral respo nses to mitigate and adapt to cli mate change , especially within the general public, is rare. Additionally, very few studies have focused on studying this relationship in rapidly developing countries such as South Korea, where greenhouse gas emissions are rising. To reinforce that evidence, after reviewing the literature meticulously in Chapter II and III, this dissertation conducts three empirical analyses (i.e., national, local, and expert levels) through Chapter IV VI by employing survey and interview data from South Korea. Finally, t his current Chapter VII summarizes the co ntent of C hapter II VI and the attainment of study objectives. It then presents lessons for policymakers and researchers respectively , and concludes with an epilogue .

PAGE 182

168 Content of Chapter II VI This section sums up two literature review chapters and three empirical analysis chapters. Chapter II . By introducing climate change, chronological global eff orts to cope with it, and significant issues negatively affecting global collaboration (e.g., complicated features, obscure leadership, and insufficient technological advancement surrounding climate change ) , t his background chapter explained why the several decade s long efforts of the international community aga inst climate change still have not come to fruition . Subsequently , policy related difficulties facing policymakers and response options in addressing climate change, along with t behavioral responses , were presented. Ch apter III . B ased on the review of theories behavioral responses to cope with climate change, this chapter identified the significant role of risk perception in that process and cast a research question, shedding light on the causal relationship among per sonal factors (experience, knowledge, and policy assessment), risk perception, and behavioral response s related to climate change. Three identified theoretical approaches ( i.e., Kates, 1971; Rogers, 1983; Grothmann & Patt, 2005 ) and previous empirical studies contribute to setting up five hypotheses and research modeling . 141 The kernel of the hypotheses is that higher risk perception affected by personal factors intensifies behavioral responses to address climate change. Chapter IV . To test the hypotheses , this first empirical analysis chapter focused on climate change risk perception and personal behaviors to mitigate climate change among individual members of the general public at the national level. Data collected from a web based 141 Simply, four of the hypotheses posit the positive relationship between experience and risk perception ( H1 ), between general knowledge and risk perception ( H2 ), between climate policy related knowledge and behavioral responses ( H3 ), and between risk perception and behavioral responses ( H5 ). The other one posits the negative relationship between policy assessment and risk perception ( H4 ).

PAGE 183

169 survey implemented in South Korea (2016, N=1,500) were analyzed quantitatively. The results of the SEM regression analysis largely supported the expectations from the literature. Even though the analysis found some personal factors, such as short term experiences (e.g., flood and drought) and the assessment of flood policy, have no relationship with climate change risk perception, most variables, such as long term climate change experiences, the understanding of the cause of climate change, a nd the assessment of drought policy significantly affect risk perception. Finally, the results showed that higher climate change risk perception, along with greater climate policy related knowledge, intensifies behavioral responses to mitigate climate chan ge . 142 This ch apter, along with the remaining empirical analysis chapters, a lso provided descriptive results and bivariate correlation coefficients among variables and concluded with policy implications, study limitations and future research directions, and wrap up. Chapter V . In the second empirical analysis chapter, the hypotheses were tested by studying drought risk perception and personal behaviors to adapt to drought among individuals at the local level . A lso, this analysis included an additional independent variable , self vulnerability perception related to limited water supply. Based on survey data from north west area of South Chungcheong Province, South Korea, during severe drought (2015, N=515), t his local level analysis employed SEM regression analysis. The results were somewhat different from what this dissertation predicted. Many personal facto rs, such as awareness of water quantity condition, ex posure to water saving information , and the familiarity with water saving subsidy policy, have a statistically significant relationship with drought risk perception or behavioral responses against drought . However, drought risk perception affected by the personal factors fails to 142 These results are derived from t he analysis of the main research model which utilizes the number of behavioral responses as a dependent variable . In the sub models which employ each behavioral response as a dependent variable, risk perception and policy related knowledge tend to promote direct (e.g., saving energy) and indirect (e.g., donations to environmental organizations) behavioral responses to mitigate climate change, respectively.

PAGE 184

170 significantly influence behavioral responses (both the index of daily action and that of device utilization associated with water saving). Self vulnerability perception has a negative relationship only with the index of water saving device utilization. Chapt er VI . The last empirical and qualitative analysis chapter tried to fill the gaps between this the results of the national and local analyses by lending an ear to expert knowledge . Based on climate change expert interviews (2018, N = 13) and survey (2018, N = 98), the following three important points were drawn. First, most hypotheses were supported by the majority of the experts both in the in terviews and survey . Second, some hypotheses (e.g., the positive relationship between e xperiences and risk perception) are likely to be actualized under certain conditions (e.g., direct, significant , and recent damage). Third, while behavioral responses to climate change, its impact is not absolute every time and everywhere . This chapter concluded that indiv es may operate differently depending on socio economic contexts recent patterns), or a complicated aspect of human nature (e.g., difficulty in changing behavioral patte rns), and that the South Korea n context helps explain some of the unexpected results . Summary of empirical evidence . Considering empirical evidence from the national, local, and expert level analyses, p olicymakers in South Korea should not assume that higher climate change risk perception will necessarily result in hig her behavioral responses to climate change. Nevertheless, the findings from this dissertation suggest that risk perception is a credible mediator in the responses to address climate change, and facilitate that process. Table 13 summarizes the results of all empirical analyses.

PAGE 185

171

PAGE 186

172 Attainment of Study Objectives T he goals of this dissertation were i mproving knowledge about the process of , advancing methodology unveiling that process , and providing viable policy implications for policymakers . This section describ es in more detail how this dissertation has achieved each of these goals. Improvement of knowledge . The findings of empirical analyses help acc umulate detailed knowledge on to clim ate change . These findings advance the literature in several ways. First, according to previous studies (e.g., Niles, Lubell, & Haden, 2013; et al ., 2005 ) , extreme weather event experiences are a significant factor influencing risk perception or behavioral responses related to climate change . However, this is not always the case, as this study showed. T he relationship between them is likely to vary in different socio economic contexts (e.g., s response capacity ). Second, individuals can perceive climate change risk based on general knowledge of climate change (e.g., cause and consequence) , which echoes existing studies (e.g., Lee et al ., 2015; Shi et al . , 2016). Further, c lima te policy related knowledge may lead individuals to take action against climate change . This articulates the role of each type of knowledge in facilitating responses . Third, a positive assessment of climate policy turns out to be one of the significant factors negatively influencing climate change risk perception in this study . This provides em pirical evidence about the assumption of people who view the government as doing their job related to climate change are less likely to feel at risk. Lastly , the positive relationship between risk perception and behavioral responses related to climate change , which many previous studies argue , is a reliable f ormula, but not an absolute

PAGE 187

173 one . T his study identifies that the impact of r isk perception can change depend ing on socio economic context s and the types of climate risk (e.g., drought) . Advancement of methodology . A qualitative approach based on expert knowledge can e nrich the finding s derived from a quantitative approach based on surveys or interviews of the general public. To identify the causal relationship surrounding risk perception and behavioral respon ses against risks, most previous studies collect ed data by c onducting surveys (or obtaining secondary datasets de rived from surveys ) and analyze d them by utilizing statistica l techniques. While the results of the se studies help expand knowledge and provide policy implications , they cannot reveal significant issues affecting th e rela tionship among variables and why some hypotheses are not supported . In overcoming this problem , e xpert interviews requiring less resources (e.g., time and money), but providing comprehensive and authoritative information from a relatively small number of i nterviews, can act as an alternative to participant observation and ordinary interviews of the general public , which consume much more resources . Importantly , t his study demonstrates the significant role of expert knowledge in elaborating the process of individua climate change. Provision of policy implications . At least two groups of policymakers directly b enefit from the policy implications of this dissertation. P olicymakers in South Korea, which is experiencing the rapid increase of GHG emissions, and those in countries with similar socio economic conditions can obtain critical clues about how to encourage the general public to voluntarily engage in be havioral responses alleviating climate change , instead of being fully responsible for climate change or forcing the general public to follow rules and regulations. Specific policy implications were already suggested in each empirical analysis. Lastly, c omprehensive lessons for policymakers are described in the following section.

PAGE 188

174 Lessons for Policymakers For scholars in professional schools, n ot only accumulating knowledge , but also illuminating practitioners are core missions in conducting research (Simon, 1967 ). A practical approach in this study ends with the following three lessons for policymakers . First, the ceaseless needed for the world to add ress climate change . This study identified that higher risk perception influenced by personal factors is likely to promote to climate change. Thus, to activate this process, policymakers should implement d iverse strategies, such as d elivering accurate and timely climate change information, communicating governmental policy against climate change , and providing behavioral guidelines , rather than just standing still . Second, policymakers strategies should be modified to fit various situation s . D iff erent contexts , such as mitigation or adaptatio n and higher or lower response capacity, are likely to require different strategies to facilitate individual behavioral responses , as are the different effects of climate change (e.g., drought and flood) . To solve difficulties in selecting the right stra tegies for a certain context, c ontinuing collaboration between the academic and the is recommended by this dissertation . Third , policymakers must remain vigilant not to miss the forest for the trees in coping with climate change. There is a possibility that strengthening the government driven climate policies , such as constructing large scale dams and tightening regulation on indus tries related to GHG emissions, behavior al responses . A ll in all, those policies, generally accompanied by the heavy financial or social burden, are in danger of making no significant progress in addressing climate change. Bueno de Mesquita (2016) warns, A ny policy lever that the government pulls to address some issue is likely to cause at least some new problems in so

PAGE 189

175 Lessons for Researchers C onsidering limitations and research directions presented in each empirical analysis chapter, t his section also provides lessons for researchers in three points. First, in fluential factors behavioral responses to address climate change vary. While this study focuses on three personal fa ctors (i.e., experience, knowledge, and policy a ssessme nt) and risk perception , t he impacts of other factors (e.g., value, media, and financial benefits ) on responses have b een identified by many previous studies. Utilizing diverse identified factors simultaneously, comparing the impact of each factor, and collecting and analyzing data derived from various contexts might be appropriate strategies for future research to develop knowledge making relat ed to climate policy. Second, depending on cross sectional survey data and single ite m based variable measurement is likely to weaken the validity of study findings . Also, while employing secondary d ata save s sources (e.g., time and money), it deprive s them of a chance to use more advanced tools in measuring variables. Unfortunately, many variables in the national and local level analyses for this study were measured b y single items from the cross sectional surveys. B ased on collaborative efforts between the academic and the s , m ore pr ecise and standardized measuremen t tools , which are able to go across lots of different countries , should be developed and shared to enhance the findings of research related to this topic. Lastly , b reaking away from traditional research methods (e.g., surveys and interviews of the general public) can facilitate the diversity of methodology in unveiling the process of individual climate change. Not only the expert interviews conducted by this study, but also other s (e.g., participant observation) are recommended. T his effort also help s solve the controversial question whether the gap between self administered behavioral responses and actual behaviors is negligible.

PAGE 190

176 Epilogue Either individually or together, climate change risk perception and individuals behavioral responses to address climate change h ave been investigated by many prominent researchers over a couple of decades . behavioral responses can play an important role in addressing it , t he behavioral response s to cope with climate change is expected to remain in the area s of both much neede d research and government strategy in the future . Although t his dissertation has made efforts to satisfy th e curiosities of the both areas across the chapters and bear s fruit to some extent, many things are still left up to future research. I really hope that more vigorous research projects make significant progress in this research field , based on the findings of this dissertation. Finally, t o help p roduce more penetrating and insightful knowledge about this complex topic , larship, which embrace s different perspectives of core stakeholders (e.g., researchers, policymakers, public manage rs, and the general public), is recommended for future studies as an appropriate research form.

PAGE 191

177 REFERENCES Abraham, C. S., Sheeran, P., Abrams, D., & Spears, R. (1994). Exploring teenagers' adaptive and maladaptive thinking in relation to the threat of HIV infection. Psychology & Health , 9 (4), Adger, W.N., Dessai, S., Goulden, M., Hulme, M., Lorenzoni, I., Nelson, D.R., Naess, L.O., Wolf, J., & Wreford, A. (2009). Are there social limits to adaptation to climate change? Climatic change , 93 Ahn, S. S. Lee, M. J., & Hwang, Y. S. (2016 ). The study on the Effect of Awareness about Climate Change and Selection Attribute toward Ski Resorts on Satisfaction and Behavior Intention: Focused on Ski Resort Visitors. Journal of Tourism and Leisure Research, 28 Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behaviour . Englewood Cliffs, NJ: Prentice Hall. Akerlof, K., Maibach, E. W., Fitzgerald, D., Cedeno, A. Y., & Neuman, A. (2013). Do people g, and if so how, and does it matter?. Global Environmental Change , 23 Akompab, D., Bi, P., Williams, S., Grant, J., Walker, I., & Augoustinos, M. (2013). Heat waves and climate change: applying the health belief model to identify predictors of risk perception and adaptive behaviours in Adelaide, Australia. International journal of environmental research and public health , 10 Aldcroft, D. H. (1983). Transport in the industrial revolution . Ma nchester: Manchester University Press. Aldy, J. E., & Stavins, R. N. (2012). Climate negotiators create an opportunity for scholars. Science , 337 Allcott, H. (2011). Social norms and energy conservation. Journal of public Economics , 95 Almalki, S. (2016). Integrating Quantitative and Qualitative Data in Mixed Methods Research -Challenges and Benefits. Journal of Education and Learning , 5 Altschuler, B., & Brownlee, M. (2016). Perceptions of climate chang e on the Island of Providencia. Local Env ironment , 21 Amsden, A. H. (1989). Asia's next giant: South Korea and late industrialization . New York, NY: Oxford University Press. Arnell, N. W. (2004). Climate change and global water resources: SRES emissions and socio economic scenarios. Global environmental change , 14

PAGE 192

178 Arrhenius, S. (1896). XXXI. On the influence of carbonic acid in the air upon the temperature of the ground. The London, Edinburgh, and Dublin Philosophical Magazine and Journal of Science , 41 Ayres, L., Kavanaugh, K., & Knafl, K. A. (2003). Within case and across case approaches to qualitative data analysis. Qualitative health research , 13 Ayton, P., & Pascoe, E. (1995). Bias in human judgement under uncertainty?. The Knowledge Eng ineering Review , 10 Azeem, M., Salfi, N. A., & Dogar, A. H. (2012). Usage of NVivo software for qualitative data analysis. Academic Research International , 2 carbon awarene ss and behavior in Tianjin, China. Energy policy , 61 Baldassare, M., & Katz, C. (1992). The personal threat of environmental problems as predictor of environmental practices. Environment and Behavior , 24 Banerjee, M., Capozzoli, M. , McSweeney, L., & Sinha, D. (1999). Beyond kappa: A review of interrater agreement measures. Canadian journal of statistics , 27 Banks, J. A. (1993). The canon debate, knowledge construction, and multicultural education. Educational researcher , 2 2 Barr, S. (2007). Factors influencing environmental attitudes and behaviors: A UK case study of household waste management. Environment and behavior , 39 Barredo, J. I. (2007). Major flood disasters in Europe: 1950 2005. Natural Haza rds , 42 148. Barrett , S . (1999). Montreal vs. Kyoto: international cooperation and the global environment . In I. Kaul, I. Grunberg & M. A. Stern (Eds.), Global Public Goods : International Cooperation in the 21st Century (pp. 192 219). New York, NY: Oxford University Press. Barrett, S. (2007). Why cooperate?: the incentive to supply global public goods . New York, NY: Oxford University Press on Demand. Bates, B., Kundzewicz, Z. W., Wu, S., & Palutikof, J. (2008). Climate change and Water: technical Paper VI . Intergovernmental Panel on Climate Change (IPCC). Battaglini, A., Barbeau, G., Bindi, M., & Badeck, F. W. (2009). Eu perceptions of climate change impact and options for adaptation. Regional Environmental Change , 9 B aum, A., Fleming, R., & Davidson, L. M. (1983). Natural disaster and technological catastrophe. Environment and Behavior , 15

PAGE 193

179 Baumgartner, F. R., Jones, B. D., & Mortensen, P. B. (2014). Punctuated equilibrium theory: Explaining stability and c hange in public policymaking. In P. A. Sabatier & C. M. Weible (Eds.), Theories of the Policy process (3rd ed., pp. Westview Press. Bellard, C., Bertelsmeier, C., Leadley, P., Thuiller, W., & Courchamp, F. (2012). Impacts of climate change on the future of biodiversity. Ecology letters , 15 Bellegarde Smith, P. (2011). A man made disaster: The earthquake of January 12, 2010 A Haitian perspective. Journal of Black Studies , 42 Bem, D. J. (1965). An experimental analysis of self persuasion. Journal of Experimental Social Psychology , 1 Beniston, M., & Diaz, H. F. (2004). The 2003 heat wave as an exa mple of summers in a greenhouse climate? Observations and climate model simulations for Basel, Switzerland. Global and planetary change , 44 Benoit, K., & Laver, M. (2007). Estimating party policy positions: Comparing expert surveys and hand coded content analysis. Electoral Studies , 26 Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin , 107 (2), 238 246. Bentler, P. M. (1995). EQS structural equations program manual. Encino, CA: Multivariate Software. Bentler, P. M., & Bonett, D. G. (1980). Significance tests and goodness of fit in the analysis of covariance structures. Psychological bulletin , 88 Berkhout, F. (2010). Reconstructing boundaries and reason in the climate debate. Global Environmental Change , 20 assessment. Climatic Change , 41 Berk, R. A., & Schulman, D. (1995). Public perceptions of global warming. Climatic Change, 29 (1), 1 33. Berse, K. (2017). Climate change from the lens of Malolos children: perception, impact and adaptation. Disaster Prevention and Management: An International Journal , 26 229. Berz, G., Kron, W., Loster, T., Rauch, E., Schimetschek, J., Schmieder , J., Siebert, A., Smolka, A., & Wirtz, A. (2001). World map of natural hazards a global view of the distribution and intensity of significant exposures. Natural hazards , 23

PAGE 194

180 Bickers, K. N., & Williams, J. T. (2001). Public policy analysis: A political economy approach . Boston, MA: Houghton Mifflin. Biernacki, P., & Waldorf, D. (1981). Snowball sampling: Problems and techniques of chain referral sampling. Sociological methods & research , 10 Biesbroek, G. R., Swart, R. J., Carter, T . R., Cowan, C., Henrichs, T., Mela, H., Morecroft, M. D., & Rey, D. (2010). Europe adapts to climate change: comparing national adaptation strategies. Global environmental change , 20 Biesbroek, G. R., Swart, R. J., & Van der Knaap, W. G. (200 9). The mitigation adaptation dichotomy and the role of spatial planning. Habitat international , 33 Bjørnæs, C. (2013). A guide to Representative Concentration Pathways. Retrieved from https://www.sei international.org/mediamanager/documents/A guide to RCPs.pdf. Blocker, T. J., & Eckberg, D. L. (1997). Gender and environmentalism: Results from the 1993 general social survey. Social Science Quarterly , 78 Bockarjova, M., & Steg, L. (2014). Can Protection Motivation Theory predict pro environmental behavior? Explaining the adoption of electric vehicles in the Netherlands. Global environmental change , 28 Bodansky, D. (2001). The history of the global climate change regime. In U. Luterbacher & D. F. Sprinz (Eds.), International relations and global climate change (pp. . Cambridge, MA: MIT Press. legitimacy. European Journal of International Law , 23 Bodansky, D. (2016). The Paris climate change agreement: a new hope?. American Journal of International Law , 110 Bogner, A., Littig, B., & Menz, W. (2009). Introduction: Expert interviews An introduction to a new methodological debate. In A. Bogner, B. Littig, & W. Menz, Interviewing experts Böhm, G., & Pfister, H. R. (2001). Mental representation of global environmental risks. In G. Böhm, J. Nerb, T. McDaniels & H. Spada (Eds.), Environmental risks: Perception, evaluation and management Bolin, B. (2007). A history of the science and politics of climate change . Cambridge, UK: Cambridge University Press. Bookchin, M. (1964). Ecology and Revolutionary Thought . New York, NY: Times Change Press.

PAGE 195

181 Booker, J. M., & McNamara, L. A. (2004). Solving black box computation problems using expert knowledge theory and methods. Reliability Engineering & System Safety , 85 (1 3), gap in environmental attitudes: the case of perceived vulnerability to risk. Social science quarterly , 78 understand global climate change?. Public under standing of science , 9 Bosello, F., Carraro, C., & De Cian, E. (2010). Climate policy and the optimal balance between mitigation, adaptation and unavoided damage. Climate Change Economics , 1 ( Bosetti, V., Carraro, C., & Tavoni, M. (2 009). Climate change mitigation strategies in fast growing countries: the benefits of early action. Energy Economics , 31 Bostrom, A., Morgan, M. G., Fischhoff, B., & Read, D. (1994). What do people know about global climate change? 1. Mental models. Risk Analysis , 14 Botzen, W. J., Aerts, J. C., & van den Bergh, J. C. (2009). Willingness of homeowners to mitigate climate risk through insurance. Ecological Economics , 68 (8 Boyatzis, R. E. (2007). Developing emotional intelligence through coaching for leadership, professional and occupational excellence. In R . Bar On , J. G. Maree & M. J. Elias ( Eds .), Educating people to be emotionally intelligent (pp. 155 168). Westport, CT: Praeger Publishers. Brandon, G., & Lewis, A. (1999). Reducing household energy consumption: a qualitative and quantitative field study. Journal of environmental psychology , 19 Brecher, J. (2015). Climate insurgency: A strategy for survival . New York, NY: Routledge. Brechin, S. R., & Bhand ari, M. (2011). Perceptions of climate change worldwide. Wiley Interdisciplinary Reviews: Climate Change , 2 Brody, S. D., Zahran, S., Vedlitz, A., & Grover, H. (2008). Examining the relationship between physical vulnerability and public perceptions of global climate change in the United States. Environment and behavior , 40 Climati c change , 90 (3), Brown, M. E., Trevino, L. K., & Harrison, D. A. (2005). Ethical leadership: a social learning perspective for construct development and testing. Organizational Behavior and Human Decision Processes , 97 (2), 117 134.

PAGE 196

182 Browne, M. W. (1984). Asymptotically distribution free methods for the analysis of covariance structures. British Journal of Mathematical and Statistical Psychology , 37 Brunnée, J., & Streck, C. (2013). The UNFCCC as a negotiation forum: towards common but more differentiated responsibilities. Climate Policy , 13 Bubeck, P., Botzen, W. J., & Aerts, J. C. (2012). A review of risk perceptions and other factors that influence flood mitigation behavior. Risk analysis , 32 Buchanan, J. & Jones, M. L. (2010). The efficacy of utilising Nvivo for interview data from the electronic gaming industry in two jurisdictions. Review of Management Innovation & Creativity , 3 Buchholz, T., Luzadis, V. A., & Volk, T. A. (2009). Sustainability criteria for bioenergy systems: results from an expert survey. Journal of cleaner production , 17 Bueno de Mesquita, E. (2016). Political Economy for Public Policy . Princeton, NJ: Princeton University Press. Campbell, J. L., Quincy, C., Osserman, J., & Pedersen, O. K. (2013). Coding in depth semistructured interviews: Problems of unitization and interc oder reliability and agreement. Sociological Methods & Research , 42 Carey, J. W. (1993). Linking qualitative and quantitative methods: I ntegrating cultural factors into public health. Qualitative Health Research , 3 Carlton, J. S., & Jacobson, S. K. (2013). Climate change and coastal environmental risk perceptions in Florida. Journal of environmental management , 130 Carlton, J. S., Mase, A. S., Knutson, C. L., Lemos, M. C., Haigh, T., Todey, D. P., & Prokopy, L. S. (2016). The effects of extreme drought on climate change beliefs, risk perceptions, and adaptation attitudes. Climatic change , 135 Catenacci, M., Verdolini, E., Bosetti, V., & Fiorese, G. (2013). Going electric: Expert survey on the future of battery technologies for electric vehicles. Energy Policy , 61 Choi, B. S., Park, K. E., & Jeon, E. C. (2014). A Study on the Reinforcement Plan f or the Local Government to Respond to the Climate Change through the Survey of Residents Consciousness Focused on the Gangnam gu . Journal of Climate Change Research , 5 Choi, H. J. (2016). The Counterstrategy of Climate Change in the Water Resources Field. In Korea Water Resources Corporation (Ed.), Water and Climate Change Seoul: Publisher Piena. (in Korean). Choi, Y. L. (2017, May 24 ). Severe drought in the South Chungcheong Province, just like in any other year. The Hankyoreh . Retrieved from http://www.hani.co.kr/arti/society/area/7960 53 .html. (i n Korean).

PAGE 197

183 Chou, C. P., & Bentler, P. M. (1995). Estimates and tests in structural equation modeling. In R. H. Hoyle (Ed.), Structural equation modeling: Concepts, issues, and applications (p p. : Sage Publications Inc. Christensen, N. S., Wood, A. W., Voisin, N., Lettenmaier, D. P., & Palmer, R. N. (2004). The effects of climate change on t he hydrology and water resources of the Colorado River basin. Climatic change , 62 Christoff, P. (2016). The promissory note: COP 21 and the Paris Climate Agreement. Environmental Politics , 25 Chungnam Institute. (2016). The 2 nd detailed plan for climate change adaptation in South Chungcheong Province (2017~2021) . Retrieved from https://www.chungnam.go.kr/dept/boa rd.do?mnu_url=/cnbbs/view.do. (i n Korean). Cialdini, R. B., Reno, R. R., & Kallgren, C. A. (1990). A focus theory of n ormative conduct: recycling the concept of norms to reduce littering in public places. Journal of personality and social psychology , 58 Application of Protection Motivation Theory. Social Marketing Quarterly , 17 Cismaru, M., Lavack, A. M., & Markewich, E. (2009). Social marketing campaigns aimed at preventing drunk driving: A review and recommendations. International Marketing Review , 26 (3), Clark, C. F., Kotchen, M. J., & Moore, M. R. (2003). Internal and external influences on pro environmental behavior: Participation in a green electricity program. Journal of environmental psychology , 23 Clayton, S., Devine Wright, P., Stern, P., Whitmarsh, L., Carrico, A., Steg, L., Swim, J., & Bonnes, M. (2015). Psychological research and global climate change. Nature Climate Change , 5 Climate Action Tracker (CAT). (2017a). Briefing paper Improving in warming outlook as India and China move ahead, but Paris Agreement gap still looms large . Retrieved from http://climateactiontracker.org/assets/publications/briefing_papers/CAT_2017 11 15_Improvement in warming outlook.pdf. Climate Action Tracker (CAT). (2017b). What is CAT? . Retrieved from http://climateactiontracker.org/. Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and Psychosocial Measurement 20 (1), 37 46. Collier, M., & Webb, R. H. (2002). Floods, droughts, and climate change . Tucson, AZ: University of Arizona Press.

PAGE 198

184 Corbett, J. B. (2005). Altruism, self interest, and the reasonable person model of environmentally responsible behavior. Science Communication , 26 Correia, F. N., Fordham, M., Saraiva, M. G., & Bernar do, F. (1998). Flood hazard assessment and management: Interface with the public. Water Resources Management , 12 (3) 227. Coxe, S., West, S. G., & Aiken, L. S. (2009). The analysis of count data: A gentle introduction to Poisson regression and its alternatives. Journal of personality assessment , 91 136. Crabbé, A., & Leroy, P. (2012). The handbook of environmental policy evaluation . Sterling, VA: Earthscan. Creswell, J. W. (2009). Research design: Qualitative, quantitativ e, and mixed methods approaches (3rd ed.). Thousand Oaks, CA: Sage. Creswell, J. W., & Plano Clark, V. L. (2007). Designing and conducting mixed methods research . Thousand Oaks, CA: Sage. Creswell , J. W. , Plano Clark , V. L. , Gutmann , M. L. , & Hanson , W. E. ( 2003 ). Advanced mixed methods research designs . In A. Tashakkori & C. Teddlie (Eds.), Handbook of mixed methods in social and behavioral research (pp. 209 240 ). Thousand Oaks, CA: Sage Publications. Crotty, J., & Lee, K. K. (2009). Was IMF imposed economic regime change i n Korea justified? The political economy of IMF intervention. Review of Radical Political Economics , 41 Curran, P. J., West, S. G . , & Finch, J. F. (1996). The robustness of test statistics to nonnormality and specification error in confirmatory factor analysis. Psychological Methods , 1 29. Dahlstrand, U., & Biel, A. (1997). Pro environmental habits: Propensity levels in behaviora l change. Journal of applied social psychology , 27 Dai, A. (2011). Drought under global warming: a review. Wiley Int erdisciplinary Reviews: Climate Change , 2 Dankelman, I. (2010). Introduction: exploring gender, environment and climate change. In I. Dankelman (Ed.), Gender and climate change: An introduction DC: Earthscan. Demski, C., Capstick, S., Pidgeon, N., Sposato, R. G., & Spence, A. (2017). Experience of extreme weather affects climate change mitigation and adaptation responses. Climatic Change , 140

PAGE 199

185 disaster experiences influence their daily behaviors regarding adaptation to climate change? A case study among young generations. Science of the Total Environment , 81 , Department for Environment, Food and Rural Affairs . (2002). Survey of public attitudes to quality of life and to the environment 2001 . London, UK: DEFRA. Department for Environment, Food and Rural Affairs . (2009). Public attitudes and behaviours towards the environment tracker survey . London, UK: DEFRA. Deser, C., Phillips, A., Bourdette, V., & Teng, H. (2012). Uncertainty in climate change projections: the role of internal variability. Climate Dynamics , 38 (3 Dessai, S., & Sims, C. (2010). Public perception of drought and climate change in southeast England. Environmental hazards , 9 57. Diggs, D. M. (1991). Drought experience and perception of climatic change among Great Plains farmers. Great Plains Research , 1 Donner, S. D., & McDaniels, J. (2013). The influence of national temperature fluctuations on opinions about clim ate change in the US since 1990. Climatic change , 118 550. Dracup, J. A. (1991). Drought monitoring. Stochastic Hydrology and Hydraulics , 5 Dracup, J. A., Lee, K. S., & Paulson, E. G. (1980). On the definition of droughts. Water resources research , 16 Dunlap, R. E., & Saad, L. (2001). Only One in Four Americans are Anxious about the Environment . Retrieved from http://www.gallup.com/poll/releases/pr010416.asp. Dunlap, R. E., & Scarce, R. (1991). The polls poll tr ends: Environment problems and protection. Public Opinion Quarterly , 55 (4), 651 672. Edwards, J. (1741). Sinners in the Hands of an Angry God. A Sermon Preached at Enfield, July 8th, 1741. Retrieved from Digital Commons at University of Nebraska Lincoln. http://digitalcommons.unl.edu/cgi/viewcontent.cgi?article=1053&context=etas. Efron, B. (1979). Bootstrap methods: Another look at the jackknife. Annals of Statistics, 7 (1), 1 26. Ellis, J. (2017, June 1). The Paris Climate Deal: What You Need to Know. The New York Times . Retrieved from https://www.nytimes.com/2017/06/01/climate/paris climate change guide.html. Erskine, H. (1972). The Polls: Pollution and Its Costs. Public Opinion Quarterly , 36 135.

PAGE 200

186 European Commission . (2017). EU Action . Retrieved f rom https://ec.europa.eu/clima/policies/international/negotiations/paris_en. Finney , S. J. , & DiStefano , C . ( 2006 ). Nonnormal and categorical data in structural equation models . In G.R. Hancock & R.O. Mueller (Eds.). Structural equation modeling: A second course (pp. ). Greenwich, CT: Information Age. Fischer, G., Shah, M., Tubiello, F. N., & Van Velhuizen, H. (2005). Socio economic and climate change impacts on agriculture: an integrated assessment, 1990 2080. Philosophical Transactions of the Royal Society B: Biological Sciences , 360 Fischhoff, B. (1985). Managing risk perceptions. Issues in Science and Technology , 2 Fischhoff, B., Lichtenstein, S., Slovic, P., Derby, S. L., & Keeney, R. L. (1981). Acceptable Risk . New Y ork, NY: Cambridge University Press. Floyd, D. L., Prentice Dunn, S., & Rogers, R. W. (2000). A meta analysis of research on protection motivation theory. Journal of applied social psychology , 30 Ford, J. D., Cameron, L., Rubis, J., Maillet, M ., Nakashima, D., Willox, A. C., & Pearce, T. (2016). Including indigenous knowledge and experie nce in IPCC assessment reports. Nature Climate Change , 6 Freedman, M., & Jaggi, B. (2011). Global warming disclosures: impact of Kyoto protocol across countries. Journal of International Financial Management & Accounting , 22 Frondel, M., Horbach, J., & Rennings, K. (2007). End of pipe or c leaner production? An empirical comparison of environmental innovation d ecisions across OECD countr ies. Business strategy and the environment , 16 Fruin, D. J., Pratt, C., & Owen, N. (1992). Protection mo perceptions of exercise. Journal of Applied Social Psychology , 22 Füssel, H. M., & Klein, R. J. (2006). Climate change vulnerability assessments: an evolution of conceptual thinking. Climatic change , 75 Gardner, G. T., & Stern, P. C. (2008). The short list: The most effective actions US households can take to curb climate change. Enviro nment: science and policy for sustainable development , 50 Gauthier, D. (1986). Morals by agreement . Oxford, UK: Clarendon Press. Gefen, D., Straub, D., & Boudreau, M. C. (2000). Structural equation modeling and regression: Guidelines for research practice. Communications of the association for information systems , 4 Giddens, A. (2011). The politics of climate change (2nd ed.). Cambridge, UK : Polity Press.

PAGE 201

187 Gifford, R., Kormos, C., & McIntyre, A. (2011). Behavioral dimensions of climate change: drivers, responses, barriers, and interventions. Wiley Interdisciplinary Reviews: Climate Change , 2 Gillroy, J. M. , & Shapiro, R. Y. (1986). The Polls: Environmental Protection. Public Opinion Quarterly , 50 Golafsh ani, N. (2003). Understanding reliability and validity in qualitative research. The qualitative report , 8 Goldberger, A. S. (1972). Structural equation methods in the social sciences. Econometrica: Journal of the Econometric Society , 40 (6), Graziano, M., & Gillingham, K. (2014). Spatial patterns of solar photovoltaic system adoption: the influence of neighbors and the built environment. Journal of Economic Geography , 15 Grothmann, T., & Patt, A. (2005). Adaptive capacit y and human cognition: the process of individual adaptation to climate change. Global Environmental Change , 15 Grothmann, T., & Reusswig, F. (2006). People at risk of flooding: why some residents take precautionary action while others do not. Natural hazards , 38 Gruber, J. (2013). Public Finance and Public Policy (4th ed.), New York, NY: Worth Publishers. Hagen, B. (2015). Public perception of climate change: Policy and communication . New York, NY: Routledge. Haines, A., Kovats, R. S., Campbell Lendrum, D., & Corvalán, C. (2006). Climate change and human health: impacts, vulnerability and public health. Public health , 120 Halady, I. R., & Rao, P. H. (2010). Does awareness to climate change lead to behavioral change? Int ernational Journal of Climate Change Strategies and Management , 2 Hallegatte, S. (2009). Strategies to adapt to an uncertain climate change. Global environmental change , 19 Hammar, H., & Löfgren, A. (2010). Explaining adoption of end of pipe solutions and clean technologies sectors in Sweden. Energy Policy , 38 Hansen, J., Sato, M., & Ruedy, R. (2012). Perception of climate change. Proceedings of the National Academy of Sciences , 109 Hanson, S., Vitek, J. D., & Hanson, P. O. (1979). Natural disaster: long range impact on human response to future disaster threats. Environment and Behavior , 11

PAGE 202

188 Harris, P.G. (2001). International Environmental Affairs and U.S. Foreign Policy. In P.G. Harris (Ed.), The Environment, International Relations, and U.S. Foreign Policy Washington, DC: Georgetown University Press. Paris pledge on climate change: inadequate and irresponsible. Journal of Environmental Studies and Sciences , 7 Heenan, A.D., & Bennis, W. (1999). Co Leaders: The Power of Great Partnerships . New York, NY: John Wiley & Sons. Heggelund, G. Asian perspective , 31 Helm, D. (2008). Climate change policy: why has so little been achieved?. Oxford review of economic policy , 24 Hilal, A. H., & Alabri, S. S. (2013). Using NVivo for data analysis in qualitative research. International Interdisciplinary Journal of Education , 2 Hooper, D., Coughlan, J. , & Mullen, M. R. (2008). Structural Equation Modelling: Guidelines for Determining M odel Fit. The Electronic Journal of Business Research Methods , 6 (1), Houghton, J. T., Callander, B. A., & Varney, S. K. (Eds.). (1992). Climate change 1992 . New York, NY: Cambridge University Press. paradigm shift and COP 21: to go beyond rhetoric. International Environmental Agreements: Politics, Law and Economics , 15 (4), Howe, P. , Markowitz, E. M. , Ming Lee, T. , Ko, C Y. , & Leiserowitz, A. ( 2013 ). Global perceptions of local temperature change . Nature Climate Change , 3 ( 4 ), 352 356 . Hox, J. J., & Bechger, T. M. (2007). An introduction to structural equation modeling. Family Science Review , 11 Hruschka, D. J., Schwartz, D., St. John, D. C., Picone Decaro, E., Jenkins, R. A., & Carey, J. W. (2004). Reliability in coding open ended data: Lessons learned from HIV behavioral research. Field methods , 16 Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional crit eria versus new alternatives. Structural equatio n modeling: a multidisciplinary journal , 6 Hu, L. T., Bentler, P. M., & Kano, Y. (1992). Can test statistics in covariance structure analysis be trusted?. Psychological bulletin , 112

PAGE 203

189 Ia cobucci, D., Saldanha, N., & Deng, X. (2007). A meditation on mediation: Evidence that structural equations models perform better than regressions. Journal of Consumer Psychology , 17 Ifinedo, P. (2012). Understanding information systems security policy compliance: An integration of the theory of planned behavior and the protection motivation theory. Computers & Security , 31 Intergovernmental Panel on Climate Change (IPCC). ( 2001). Climate Change 2001: Impacts, Adaptation, and Vulnerability. Summary for Policymakers . Cambridge, UK: Cambridge University Press. Intergovernmental Panel on Climate Change (IPCC). (2014a). IPCC . Retrieved from http://www.ersilia.org/canvi_climatic/documents/IPPC/IPCC_Whoiswho.pdf. Intergovernmental Panel on Climate Change (IPCC). (2014b). Fifth Assessment Report: Climate Change 2014. Synthesis Report. Geneva: IPCC. Intergovernmental Panel on Climate Change (IPCC ). (2014c). Climate Change 2014 Impacts, Adaptation and Vulnerability . New York, NY: Cambridge University Press. Intergovernmental Panel on Climate Change (IPCC) . (2017). IPCC Organization . Retrieved from http://www.ipcc.ch/organization/organization_structure.shtml . International Telecommunication Union. (2017). ICT Development Index 2017. Retrieved from http://www.itu.int/net4/ITU D/idi/2017/index.html. Ito, T. (2007). Asian currency crisis and the International Monetary Fund, 10 years l ater: Overview. Asian Economic Policy Review , 2 Jacobsson, S., & Lauber, V. (2006). The politics and policy of energy system transformation explaining the German diffusion of renewable energy technology. En ergy policy , 34 (3), M., Parry, M., Peters, I., Ravetz, J., & Rotmans, J. (1997). Procedural leadership in climate policy: a European task. Global Envir onmental Change , 7 Jaeger, C., Dürrenberger, G., Kastenholz, H., & Truffer, B. (1993). Determinants of environmental action with regard to climatic change. Climatic Change , 23 Jentsch, A., Kreyling, J., & Beierkuhnlein, C. (2007). A new generation of climate change experiments: events, not trends. Frontiers in Ecology and the Environment , 5 374. Johnson E. J., & Tversky, A. (1983). Affect, generalization, and the perception of risk. Journal of Personality and Social Psychol ogy , 45 (1)

PAGE 204

190 Johnson, R. B., & Onwuegbuzie, A. J. (2004). Mixed methods research: A research paradigm whose time has come. Educational researcher , 33 Johnson, R. B., Onwuegbuzie, A. J., & Turner, L. A. (2007). Towar d a definition of mixed methods research. Journal of mixed methods research , 1 Jones, B. D., & Baumgartner, F. R. (2012). From there to here: Punctuated equilibrium to the general punctuation thesis to a theory of government information proces sing. Policy Studies Journal , 40 Jones, R. N. (2001). An environmental risk assessment/management framework for climate change impact assessments. Natural hazards , 23 Jones, R. N., Dettmann, P., Park, G., Rogers, M., & White, T. (2007). The relationship between adaptation and mitigation in managing climate change risks: a regional response from North Central Victoria, Australia. Mitigation and Adaptation Strategies for Global Change , 12 Jonkman, S. N. (2005). Global p erspectives on loss of human life caused by floods. Natural hazards , 34 Jöreskog, K. G., & Sörbom, D. (1984). USREL VI user's guide (3rd ed.). Mooresville, IL: Scientific Software. Joshi, S., Jasra, W. A., Ismail, M., Shrestha, R. M., Yi, S. L perceptions of and responses to climate change in northern Pakistan. Environmental management , 52 Kahan, D. M., Peters, E., Wittlin, M., Slovic, P., Ouellette, L. L., Braman, D., & Mandel, G. (2012). The polarizing impact of science literacy and numeracy on perceived climate change risks. Nature climate change , 2 Kaiser, H. F. (1974). An index of factorial simplicity. Psychometrika , 39 Kane, S., & Shogren, J. F. (2000). Linking adaptation an d mitigation in climate change policy. Climatic Change , 45 Kane, S., & Yohe, G. (2000). Societal adaptation to climate variability and chan ge: an introduction. In S. Kane & G. Yohe (Eds.), Societal Adaptation to Climate Variability and Change ( Dordrecht : Springer Netherlands. Karlsson, C., Parker, C., Hjerpe, M., & Linnér, B. O. (2011). Looking for leaders: Perceptions of climate change leadership among climate change negotiation participants. Global Environmental Politics , 11 (1), Karp, D. G. (1996). Values and their effect on pro environmental behavior. Environment and behavior , 28

PAGE 205

191 Kates, R. W. (1971). Natural hazard in human ecological perspective: hypotheses and models. Economic Geography , 47 Kaul, I., Conceicao, P., Le Goulven, K., & Mendoza, R. U. (Eds.). (2003). Providing global public goods: managing globalization . New York, NY: Oxford University Press. Kaul, I., Grunberg, I., & Stern, M. A. (1999). Global public goods . New York, NY: Oxford University Press. Kearney, A. R. (1994). Understanding global change: a cognitive perspective on communicating through stories. Climatic Change , 27 Keller, C., Siegrist, M., & Gutscher, H. (2006). The role of the affect and availability heuri stics in risk communication. Risk analysis , 26 Kellstedt, P. M., Zahran, S., & Vedlitz, A. (2008). Personal efficacy, the information e nvironment, and attitudes toward global warming and climate change in the United States. Risk Analysis , 28 environmental behavior under drought: Application of protection motivation theory. Journal of Arid Environments , 127 136. Kiesler, C. A. (1971). The psychology of commitment: Experiments l inking behavior to belief . New York, NY: Academic Press. Climate Change in Korea. Journal of Fisheries Business Administration , 45 (in Korean). Kim, D. S. (2 017, June 7). Rain not enough to quench ongoing drought. The Korea Herald . Retrieved from http://www.koreaherald.com/view.php?ud=20170607000739. Kim, G. C., Chang, W. J., & Won, H. P. (2015, October 11). Drought led South Korea to cut water supply in some towns. The Maeil Business News Korea . Retrieved from http://pulsenews.co.kr/view.php?year=2015&no=968843. Kim, J. I. (2016). A study on the introduction of water demand response system in preparation for drought. In Korea Water Resources Corporation (Ed.), Water and Climate Change Kim, J. O., & Kim, N. J. (20 13). An Analysis of Structural Relationship between Climate application of the Extended Theory of Planned Behavior. Journal of Tourism and Leisure Research, 25 Korean).

PAGE 206

192 Kim, O., Kim, B. B., Lee, J. H. (2014). Awareness of Climate Change among Some Residents in Chungnam Province. Journal of Korean Society for Environmental Education , 27 82. (in Korean). Kim, P. S. & Moon, M. J . (2002). Current Public Sector Reform in Korea: New Public Management in Practice. Journal of Comparative Asian Development , 1 (1), Kim, S., Jeong, S. H., & Hwang, Y. (2013). Predictors of pro environmental behaviors of American and Korean students: The application of the theory of reasoned action and protection motivation theory. Science Communication , 35 Kim, S. Y. (2011). Public perceptions of climate change and support for climate policies in Asia: evidence from recent polls. The Journal of Asian Studies , 70 Kitous, A., Keramidas, K., Vandyck, T., & Saveyn, B. (2016). GECO 2016 Global Energy and Climate Outlook Road from Paris . Sevilla: European Commission. Klein, R. J., Schipper, E. L. F., & Dessai, S. (2005). Integra ting mitigation and adaptation into climate and development policy: three research questions. Environmental science & policy , 8 Kline, R. B. (2015). Principles and practice of structural equation modeling (4th ed.). New York, NY: Guilford publ ications. Koh, J. K., & Yi, W. P. (2016). A Study on Fostering Climate Change Adaptation Action. Journal of Environmental Policy and Administration , 24 Kollmuss, A., & Agyeman, J. (2002). Mind the gap: why do people act environmen tally and what are the barriers to pro environmental behavior?. Environmental education research , 8 Kolstad, C. D. (2000). Environmental Economics . New York, NY: Oxford University Press. Korea Water Resources Association (KWRA). (2018). About KWRA . Retrieved from http://www.kwra.or.kr/eng/main/index.htm. Korea Water Resources Corporation ( K water). (2013). Review of water supply in the South Chungcheong Province . Daejeon: K water. (in Korean). Korea Water Resources Corporation ( K water). (2016 a ). A Study on the Effects of Water Resource Development during the Korea Development Period . Daejeon: K water Institute. (in Korean). Korea Water Resources Corporation ( K water) . (2016b). A study on Water Resource Strategies for Climate Change . Daejeon: K water Institute. (in Korean).

PAGE 207

193 Korea Water Resources Corporation (K water). (2018). About K water. Retrieved from http://english.kwater.or.kr/eng/about/sub02/kwaterPage.do?s_mid=1099 and http://english.kwater.or.kr/eng/about/sub01/massagePage.do?s_mid=109 7. Korea Water Resources Corporation Institute (K water institute). (2018). About K water Institute. Retrieved from https://kiwe.kwater.or.kr/english/guide/sub01/planPage . do?s_mid=1238. Krippendorff, K. (2004). Content Analysis: An Introduction to Its Met hodology (2nd ed). Thousand Oaks, CA: Sage. Krueger, T., Page, T., Hubacek, K., Smith, L., & Hiscock, K. (2012). The role of expert opinion in environmental modelling. Environmental Modelling & Software , 36 Kuhnert, P. M., Martin, T. G., & Griffiths, S. P. (2010). A guide to eliciting and using expert knowledge in Bayesian ecological models. Ecology letters , 13 Kuhn, T. S. (1962). The structure of scientific revolutions . Chicago, IL: University of Chicago Press. Kvale, S. (1996) . InterViews: An introduction to qualitative research interviewing . Thousand Oaks, CA: Sage. Kvale, S. (2003). The psychoanalytic interview as inspiration for qualitative researc h. In P. M. Camic, J. E. Rhodes & L. Yardley (Eds.), Qualitative research in psychology: Expanding perspectives in methodology and design (pp. 275 297). Washington, DC: American Psychological Association. Lakatos, I., & Musgrave, A. (1970). Criticism and growth of knowledge . Cambridge: Cambridge University press. LaMoreaux, P. E. (2001). Historical development. In P. E. LaMoreaux & J. T. Tanner (Eds.), Springs and bottled waters of the world: Ancient history, source, occurrence, quality and use Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics , 33 Larson, J. F., & Park, J. (2014). From developmental to network state: Government restructuring and ICT led innovation in Korea. Telecommunications Policy , 38 59. Latané, B., Williams, K., & Harkins, S. (1979). Many hands make light the work: The causes and consequences of social loafing. Journal of personality and social psychology , 37 (6), Laukkonen, J., Blanco, P. K., Lenhart, J., Keiner, M., Cavric, B., & Kinuthia Njenga, C. (2009). Combining climate change adaptation and mitigation measures at the local level. Habitat International , 33

PAGE 208

194 and barriers to adaptation: lessons learned from an exploratory study in V ietnam. Mitigation and adaptation strategies for global change , 19 Lee, M. J., Hwang, Y. S., & Seo, W. S. (2015). Exploring the Differences in Travel Risk and sfaction and Behavior Intention on Jeju Tourism According to the Perception on Climate Change. Journal of Tourism Studies , 27 Lee, S. M., & Ungson, G. R. (2008). Towards a theory of synchronous technological assimilation: The case s Internet economy. Journal of World Business , 43 (3), Lee, T. M., Markowitz, E. M., Howe, P. D., Ko, C. Y., & Leiserowitz, A. A. (2015). Predictors of public climate change awareness and risk perception around the world. Nature climate c hange , 5 Leiserowitz, A. (2004). Day after tomorrow: study of climate change risk perception. Environment: Science and Policy for Sustainable Development , 46 Leiserowitz, A. (2005). American risk perceptions: Is climate change dangerous?. Risk analysis , 25 Leiserowitz, A. (2006). Climate change risk perception and policy preferences: The role of affect, imagery, and values. Climatic change , 77 Leiserowitz, A. (2007). International public opinion, percep tion, and understanding of global climate change. Human development report 2007/ 2008 Lemieux, C. J., Gray, P. A., Douglas, A. G., Nielsen, G., & Pearson, D. (2014). From science to policy: The making of a watershed scale climate change adaptation st rategy. Environmental Science & Policy , 42 Washington Quarterly , 31 Lewis, J. I., & Wiser, R. H. (2007). Fostering a renewable energy technology industry: An international comparison of wind industry policy support mechanisms. Energy policy , 35 Libakova, N. M., & Sertakova, E. A. (2015). The method of expert interview as an effective research procedure of studying the indi genous peoples of the north. Journal of Siberian Federal University. Humanities & Social Sciences, 1 Lim, B., Spanger Siegfried, E., Burton, I., Malone, E., & Huq, S. (2004). Adaptation policy frameworks for climate change: developing strategi es, policies and measures . Cambridge, UK: Cambridge University Press.

PAGE 209

195 Lindell, M. K., & Hwang, S. N. (2008). Households' perceived personal risk and responses in a multihazard environment. Risk Analysis , 28 Linsley Jr, R. K., Kohler, M. A., & Paulhus, J. L. (1975). Hydrology for engineers . New York, NY: McGraw Hill. Liu, J. (2011). The Cancun Agreements. Environmental Law Review , 13 Liu, Z., Smith, W. J., & Safi, A. S. (2014). Rancher and farmer perceptions of climate change in Nevada, USA. Climatic Change , 122 (1 Li, Y., Johnson, E. J., & Zaval, L. (2011). Local warming: Daily temperature change influences belief in global warming. Psychological science , 22 Lloyd Hughes, B. (2014). The impracticality of a universal drought definition. Theoretical and applied climatology , 117 Lo, A. Y. (2013). The role of social norms in climate adaptation: Mediating risk perception and flood insurance purchase. Global Environmental Change , 23 Lobell, D. B., Burke, M. B., Tebaldi, C., Mastrandrea, M. D., Falcon, W. P., & Naylor, R. L. (2008). Prioritizing climate change adaptation needs for food security in 2030. Science , 319 Lombardi, D., & Sinatra, G. M. (2012). College studen human induced climate change. Research in Science Education , 42 Lombard, M., Snyder Duch, J., & Bracken, C. C. (2002). Content analysis in mass communication: Assessment and reporting of intercoder reliability. Human communication research , 28 Lorenzoni, I., & Pidgeon, N. F. (2006). Public views on climate change: European and USA perspectives. Climatic change , 77 Lowe, T., Brown, K., Dessai, S., de França Doria, M., Haynes, K., & Vincent, K. (2006). Does tomorrow ever come? Disaster narrative and public perceptions of climate change. Public understanding of science , 15 Lubell, M., Zahran, S., & Vedlitz, A. (2007). Collective action and citizen responses to global warming. Political Behavior , 29 Lund, H., & Mathiesen, B. V. (2009). Energy system analysis of 100% renewable energy systems The case of Denmark in years 2030 and 2050. Energy , 34 Lutsey, N., s bott om up climate change mitigation policy. Energy policy , 36

PAGE 210

196 Mainieri, T., Barnett, E. G., Valdero, T. R., Unipan, J. B., & Oskamp, S. (1997). Green buying: The influence of environmental concern on consumer behavior. Journal of social psychology , 137 Malka, A., Krosnick, J. A., & Langer, G. (2009). The association of knowledge with concern about global warming: Trusted information sources shape public thinking. Risk Analysis , 29 Marketlink. (2017). Panel Infrastructure. Retrieved from http://www.marketlink.co.kr/panel/panelStatus.php. Marquart Pyatt, S. T., McCright, A. M., Dietz, T., & Dunlap, R. E. (2014). Politics eclipses climate extremes for climate change perceptions. Global Environmental Change , 29 , Martin, T. G., Burgman, M. A., Fidler, F., Kuhnert, P. M., LOW CHOY, S. A. M. A. N. T. H. A., McBride, M., & Mengersen, K. (2012). Eliciting expert knowledge in conservation science. Conservation Biology , 26 Martin, W. E., Martin, I. M., & Kent, B. (2009). The role of risk per ceptions in the risk mitigation process: the case of wildfire in high risk communities. Journal of environmental management , 91 Mase, A. S., Gramig, B. M., & Prokopy, L. S. (2017). Climate change beliefs, risk perceptions, and adaptation behavior among Midwestern US crop farmers. Climate Risk Management , 15 Masud, M. M., Al Amin, A. Q., Junsheng, H., Ahmed, F., Yahaya, S. R., Akhtar, R., & Banna, H. (2016). Climate change issue and theory of planned behaviour: relationship by empirical evidence. Journal of Cleaner Production , 113 Mayer, D.M., Kuenzi, M., Greenbaum, R., Bardes, M. , & Salvador, R. (2009). How low does ethical leadership flow? Test of a trickle down model. Organizational Behavi or and Human Decision Processes , 108 (1), 1 13. McBride, M. F., & Burgman, M. A. (2012). What is expert knowledge, how is such knowledge gathered, and how do we use it to address questions in landscape ecolog y?. In A. H. Perera, C. A. Drew & C. J. Johnson (Eds.), Expert knowledge and its application in landscape ecology McCrae, R. R. (1984). Situational determinants of coping responses: Loss, threat, and challenge. Journal of personality and Social Psychol ogy , 46 McCright, A. M. (2010). The effects of gender on climate change knowledge and concern in the American public. Population and Environment , 32

PAGE 211

197 McCright, A. M., & Dunlap, R. E. (2011). The politicization of climate change and p olarization 2010. The Sociological Quarterly , 52 McDonald, R. P., & Ho, M. H. R. (2002). Principles and practice in reporting structural equation analyses. Psychological methods , 7 (1), McMichael, A. J., Powles, J. W., Butler, C. D., & Uauy, R. (2007). Food, livestock production, energy, climate change, and health. The lancet , 370 McMichael, A. J., Woodruff, R. E., & Hales, S. (2006). Climate c hange and human heal th: present and future risks. The Lancet , 367 Meehl, G. A., Zwiers, F., Evans, J., Knutson, T., Mearns, L., & Whetton, P. (2000). Trends in extreme weather and climate events: issues related to modeling extremes in projections of future cli mate change. Bulletin of the American Meteorological Society , 81 Mehran, A., Mazdiyasni, O., & AghaKouchak, A. (2015). A hybrid framework for assessing socioeconomic drought: Linking climate variability, local resilience, and demand. Journal o f Geophysical Research: Atmospheres , 120 Meijerink, S., & Stiller, S. (2013). What kind of leadership do we need for climate adaptation? A framework for analyzing leadership objectives, functions, and tasks in climate change adaptation. Environment and Planning C: Government and Policy , 31 Mendelsohn, R., Dinar, A., & Williams, L. (2006). The distributional impact of climate change on rich and poor countries. Environment and Development Economics , 11 Meuser, M., & Nagel, U. (2009). The expert interview and changes in knowledge production. In A. Bogner, B. Littig, & W. Menz, Interviewing experts Palgrave Macmillan. Meyerson, F. A. (1998). Population, Carbon Emissions, and Global Warming: The Forgotten Relationship at Kyoto. Population and Development Review , 24 Miceli, R., Sotgiu, I., & Settanni, M. (2008). Disaster preparedness and perception of flood risk: A study in an alpine valley in Italy. Journal of environmental psycho logy , 28 (2), 173. Miles, M., Huberman, A., & Saldaña, J. (2014). Qualitative data analysis: A methods sourcebook (3nd ed. ). Thousand Oaks, CA: Sage Publications Inc. Milfont, T. L. (2012). The interplay between knowledge, perceived efficacy, and concern about global warming and climate change: a one year longitudinal study. Risk Analysis , 32 (6), Millar, C. I., Stephenson, N. L., & Stephens, S. L. (2007). Climate change and forests of the future: managing in the face of uncertainty. Ecological applications , 17

PAGE 212

198 Milne, S., Sheeran, P., & Orbell, S. (2000). Prediction and intervention in health related behavior: A meta analytic review of protection motivation theory. Journal of Applied Social Psychology , 30 Mi nistry of Education. (2018). Distribution of population by Education. Retrieved from http://www.index.go.kr/potal/main/EachDtlPageDetail.do?idx_cd=1530. (in Korean). Ministry of Environment. (2014). 2013 national consciousness survey on environmental conse rvation. Retrieved from http://www.me.go.kr/home/web/public_info/read.do ; jsessionid=h9RATBW1vUIsLAZh4nw9FLL6IaafPhtOW3c4Ma2F2EXy6yM4ObZ1UZh kqfSvJ9Xk.meweb1vhost_servlet_engine1?pagerOffset=0&maxPageItems=10&maxInde xPages=10&searchKey=&searchValue=&menuId= 10357&orgCd=&condition.publicInfo MasterId=8&publicInfoId=453&menuId=10357. (in Korean). Ministry o f Government Legislation . (2015a). Act on the Promotion of Saving and Recycling of Resources. Retrieved from http://www.law.go.kr/eng/engLsSc.do?menuId=1& query=Act +on+the+Promotion+of+Saving+and+Recycling+of+Resources&x=0&y=0#liBgcolor0. Ministry o f Government Legislation . (2015b). Korea Water Resources Corporation Act . Retrieved from http://www.law.go.kr/LSW/eng/engLsSc.do?menuId=2§ion= lawNm&query=KO REA+WATER+RESOURCES+CORPORATION+ACT&x=35&y=2 5#liBgcolor2 . Ministry o f Government Legislation . ( 2016). Water Supply and Waterworks Installation Act. Retrieved from http://www.law.go.kr/LSW/eng/engLsSc.do?menuId=2§ion=lawNm &query=Water+Supply+and+Waterw orks+Installation+Act&x=22&y=56#liBgcolor18 . Ministry o f Government Legislation . (2017 a ). Storm and Flood Insurance Act. Retrieved from http://www.law.go.kr/LSW/eng/engLsSc.do?menuId=2§ion=lawNm &query=Storm+and+Flood+Insurance+Act&x=42&y=39#liBgcolor0 . Ministry of Government Legislation. (2017b). Local Autonomy Act. Retrieved from http://www.law.go.kr/LSW/eng/engLsSc.do?menuId=2§ion=lawNm&query=LOCA L+AUTONOMY+ACT&x=18&y=18#liBgcolor15. Ministry of Infrastructure and Transport (MOLIT). (2015, October 30). Starting to build a water conduit from the Baekje bo to the Boryeong dam . Retrieved from http://www.molit.go.kr/USR/N EWS/m_71/dtl.jsp?id=95076494. (i n Korean). Ministry of Infrastructure and Transport (MOLIT) and Korea Water Resources Corporation (K water) . (2017). Water and Sustainable Development . Retrieved from http://www.kwater . or.kr/news/sub03/sub01/landList.do?s_mid=232&seq=212. (in Korean). Ministry of the Interior and Safety (MOIS). (2017). Statistical Yearbook of Natural Disaster 2017 . Retrieved from http://www.safekorea.go.kr/idsiSFK/neo/sfk/cs/csc/bbs_conf.jsp ? bbs _no=26&emgPage=Y&menuSeq=735. (i n Korean)

PAGE 213

199 Min, W. K., Kim, H. J. , & Kim, N. J. (2015). The Effects of Consumers Risk Perceptions to Climate Change on Emotional Consumption value and Purchasing Intent towards Environment friendly Tourism Products. Journal of Tourism and Leisure Research , 27 Mirza, M. M. Q. (2003). Climate change and extreme weather events: can developing countries adapt?. Climate policy , 3 Mitchell, J. K. (1974). Natural hazards research. Perspectives on Environment, Asso. Am. Geographers Pub Montgomery, D. C. , Peck, E. A., & Vining, G. G. (2012). Introduction to linear regression analysis (5th ed.). Hoboken, NJ: John Wiley & Sons. Moss, R., Babiker, M., Brinkman, S., Calvo, E., Carter, T., Edmonds, J., ... & Zurek, M. (2008). Towards New Scenarios for the Anal ysis of Emissions, Climate Change, Impacts, and Response Strategies . Geneva: Intergovernmental Panel on Climate Change . Mulilis, J. P., & Lippa, R. (1990). Behavioral change in earthquake preparedness due to negative threat appeals: A test of protection mo tivation theory. Journal of Applied Social Psychology , 20 Musinguzi, L., Efitre, J., Odongkara, K., Ogutu Ohwayo, R., Muyodi, F., Natugonza, V., ... & and adaptation strategies in environmental change hotspots: a case of Lake Wamala, Uganda. Environment, development and sustainability , 18 National Aeronautics and Space Administration (NASA). (2011). What Are Climate and Climate Change?, Retrieved from https://www.nasa.gov/audience/forstudents/5 8/features/nasa knows/what is climate change 58.html. National Drought Information Analysis Center (2018). 2013~2018 Sustained Drought Analysis & Assessment Report . Retrieved from http://drought.kwater.or.kr/menu/m80/m83_view . do?page=1&manage_no=1 44. (i n Korean) . National Information Society Agency. (2017). 2016 National Informatization White Paper . Retrieved from http://www.nia.or.kr/site/nia_kor/ex/bbs/View.do?cbIdx = 44086&bcIdx=18304&parentSeq=18304. (in Korean). Neuendorf, K. A. (2017). The content analysis guidebook (2nd ed.) Thousand Oaks, CA: Sage. Nevitt, J., & Hancock, G. R. (2001). Performance of bootstrapping approaches to model test statistics and parameter standard error estimation in structural equation modeling. Structural Equation Modeling , 8 (3), Nhamo, G. (2009). Co leadership in climate change: An agenda to 2013 and beyond. Politikon , 36

PAGE 214

200 Niles, M. T., Lubell, M., & Haden, V. R. (2013). Perceptions and responses to climate policy risks among California farmers. Global environmental change , 23 Nolan, J. M., Schul tz, P. W., Cialdini, R. B., Goldstein, N. J., & Griskevicius, V. (2008). Normative social influence is underdetected. Personality and social psychology bulletin , 34 Nordhaus, W. D. (1991). To slow or not to slow: the economics of the greenhouse effect. The economic journal , 101 Northouse, P.G. (2016). Leadership: Theory and practice (7th ed.). Thousand Oaks, CA: Sage Publications. Novotny, V., Ahern, J., & Brown, P. (2010). Water centric sustainable communities: planning , retrofitting and building the next urban environment . Hoboken, NJ: John Wiley & Sons. Nyong, A. (2009). Climate change impacts in the developing world: implications for sustainable development. In L. Brainard, A . Jones & N. Purvis (Eds.), Climate Change and Global Poverty: a billion lives in the balance Institution Press. Oberthür, S., & Roche Kelly, C. (2008). EU leadership in internation al climate policy: achievements and challenges. The International Spectator , 43 beliefs, and willingness to address climate change. Risk analysis , 19 (2005). Feeling at risk matters: water managers and the decision to use forecasts. Risk Analysis , 25 1275. Withdrawal from Climate Change Agreements: Implications for Carbon Emissions Reduction. Journal of Human Ecology , 41 Oakley, J. E., & Rakow, T. (2006). Uncertai probabilities . Ch ichester, U.K.: John Wiley & Sons. Olson, J. M., & Stone, J. (2005). The influence of behavior on attitudes. In D. Albarracin, B. T. Johnson & M. P. Zanna (Eds.), The handbook of attitudes Lawrence Erlbaum. Opensurvey. (2017). Panel Information. Retrieved from http://www.opensurvey.co.kr/panel/. Opiyo, F., Wasonga, O. V., Nyangito, M. M., Mureithi, S. M., Obando, J., & Munang, R. (2016). Determinants of perceptions of climate change and adaptation among Turkana pastoralists in n orthwestern Kenya. Climate and Development , 8

PAGE 215

201 Oreskes, N. (2004). The scientific consensus on climate change. Science , 306 Organization for Economic Cooperation and Development ( OECD). (2016). Total fixed and wireless broadband subscription by country . Retrieved from http://www.oecd.org/sti/ broadband/oecdbroadbandportal.htm. Organization for Economic Cooperation and Development ( OECD). (2017). Air and GHG emissions (indicator). Retrieved from doi: 10.1787/93d10cf7 en. Osberghaus, D., Dannenberg, A., Mennel, T., & Sturm, B. (2010). The role of the government in adaptation to climate change. Environment and pla nning C: government and policy, 28 (5), Ostrom, E., Burger, J., Field, C. B., Norgaard, R. B., & Policans ky, D. (1999). Revisiting the commons: local lessons, global challenges. science , 284 Ostrom, E., Gardner, R., & Walker, J. (1994). Rules, games, and common pool resources . Ann Arbor, MI: University of Michigan Press. Pachauri, R. K., Allen , M. R., Barros, V. R., Broome, J., Cramer, W., Christ, R., ... & Dubash, N. K. (2014). Climate change 2014: synthesis report. Contribution of Working Groups I, II and III to the fifth assessment report of the Intergovernmental Panel on Climate Change . Gen eva, Switzerland: IPCC. Palmer, W. C. (1965). Meteorological drought (Vol. 30). Washington, DC: US Department of Commerce, Weather Bureau. Papoulis, D., Kaika, D., Bampatsou, C., & Zervas, E. (2015). Public Perception of Climate Change in a Period of Econo mic Crisis. Climate , 3 Perception to Climate Change in Korea. Journal of Climate Change Research , 5 46. (in Korean). Park, H. (2011). Man made disasters: A cross national analysis. International Business Review , 20 Park, Y. J. (2018, October 31). The percentages of female professors in universities are only 26%. Skyedaily . Retrieved from http://www.skyedaily.com/news/news_spot.html?ID=78536 . Parker, C. F., Karlsson, C., Hjerpe, M., & Linnér, B. O. (2012). Fragmented climate change leadership: making sense of the ambiguous outcome of COP 15. Environmental Politics, 21 Patchen, M. (2006). Public attitudes and behavior about climate change . West Lafayette, IN: PCCRC Outreach Publication.

PAGE 216

202 Patton, M. Q. (2002). Qualitative Research and Evaluation Methods (3rd ed.). Thousand Oaks, CA: Sage Publications Inc. Patz, J. A., Campbell Lendrum, D., Holloway, T., & Foley, J. A. (20 05). Impact of regional climate change on human health. Nature , 438 Pearce, D. W., Cline, W. R., Achanta, A. N., Fankhauser, S., Pachauri, R. K., Tol, R. S., & Vellinga, P. (1996). The social costs of climate change: greenhouse damage and the benefits of control. In J. P. Bruce, H. Lee & E. F. Haites (Eds.), Climate change 1995: Economic and social dimensions of climate change Cambridge, UK: Cambr idge University Press. Peduzzi, P., Dao, Q. H., Herold, C., & Mouton, F. (2003). Global Risk and Vulnerability Index Trends per Year (GRAVITY) Phase III: Drought analysis . Retrieved from https://archive ouverte.unige.ch/unige:32354. Pereira, L. S., Cordery, I., & Iacovides, I. (2009). Coping with water scarcity: Addressing the challenges . Dordrech: S pringer . Perry, R. W., Lindell, M. K., & Tierney, K. J. (2001). Facing the un expected: Disaster preparedness and response in the United States . Washington, DC: Joseph Henry Press. Pidgeon, N. (1998). Risk assessment, risk values and the social science programme: why we do need risk perception research. Reliability Engineering & System Safety , 59 Pidgeon, N. , & Gregory, R. (2004). Judgeme nt, decision making and public policy. In D. Koehler & N. Harvey (Eds.), Blackwell Handbook of Judgement and Decision Making ( pp. 604 623). Oxford, UK: Blackwell Publishing. Pielke, R. A. (1994). Scientific information and global change policymaking [1]. Climatic Change , 28 Pielke, R. A. (1998). Rethinking the role of adaptation in climate policy. Global environmental change , 8 Piwowarczyk, J., Hansson, A., Hjerpe, M., Chubarenko, B., & Karmanov, K. (2012). Climate change in the B altic Sea region: A cross country analysis of institutional stakeholder perceptions. Ambio , 41 Plumer, B. (2017, September 20). How Can U.S. States Fight Climate Change if Trump Quits the Paris Accord? The New York Times . Retrieved from https://www.nytimes.com/ 2017/09/20/climate/paris climate accord trump.html. Poussin, J. K., Botzen, W. W., & Aerts, J. C. (2014). Factors of influence on flood damage mitigation behaviour by households. Environmental Science & Policy , 40

PAGE 217

203 Pulwarty, R. S., & Riebsame, W. E. (1997). The political ecology of vulnerability to hurricane related hazards . In H.F. Diaz & R.S. Pulwarty (Eds.), Hurricanes . Climate and socioeconomic impacts ( pp. 185 214) . New York, NY: Springer Verlag. Rainear, A. M., & Christensen, J. L. (2017). Protection Motivation Theory as an Explanatory Framework for Proenvironmental Behavioral Intentions. Communication Research Reports , 34 Rainey, H. G. (1994). On paradigms, progress, and prospects for public managemen t. Journal of Public Administration Research and Theory: J PART , 4 Ray, L. (1999). Measuring party orientations towards European integration: Results from an expert survey. European Journal of Political Research , 36 Read, D., Bostrom, A., Morgan, M. G., Fischhoff, B., & Smuts, T. (1994). What do people know about global climate change? 2. Survey studies of educated laypeople. Risk Analysis , 14 (6), Rebetez, M. (1996). Public expectation as an element of human perception of climate change. Climatic Change , 32 Regan, M. D. (2017, November 11). U.S. cities, states pledge support for climate accord. Public Broadcasting Service (PBS) . Retrieved from https://www.pbs.org/newshour/nation/u s cities states pledge support for climate accord. Revi, A. (2008). Climate change risk: an adaptation and mitigation agenda for Indian cities. Environment and Urbanization , 20 Riccucci, N. M. (2010). Public administration: Traditions of inquiry and philosophies of knowledge . Washington, DC: Georgetown University Press. Rippetoe, P. A., & Rogers, R. W. (1987). Effects of components of protection motivation theory on adaptive and maladaptive coping with a health threat. Journal of personality and social psychology , 52 Roco, L., Engler, A., Bravo Ureta, B. E., & Jara climate change in mediterranean Chile. Regional Environmental Change , 15 879. R ogelj, J., Den Elzen, M., Höhne, N., Fransen, T., Fekete, H., Winkler, H., Schaeffer, R., Sha, F., Riahi, K., & Meinshausen, M. (2016). Paris Agreement climate proposals need a boost to keep warming well below 2 C. Nature , 534 Rogers, R. W. (1975). A protection motivation theory of fear appeals and attitude change . The Journal of Psychology , 91

PAGE 218

204 Rogers, R. W. (1983). Cognitive and physiological processes in fear appeals and attitude change: A revised theory of protection motivatio n. In B. L. Cacioppo , & L. L. Petty (Eds.), Social Psychophysiology: A Sourcebook (pp. 153 176). London, UK: Guilford. Rosenzweig, C., Iglesias, A., Yang, X. B., Epstein, P. R., & Chivian, E. (2001). Climate change and extreme weather events; implications for food production, plant diseases, and pests. Global change & human health , 2 Rosenzweig, C., & Parry, M. L. (1994). Potential impact of clima te change on world food supply. Nature , 367 Rubin, H.J. , & Rubin, I.S. (2012). Qualitative Interviewing: The Art of Hearing Data (3rd ed. ). Thousand Oaks, CA: Sage Publications. Ruddell, D., Harlan, S. L., Grossman Clarke, S., & Chowell, G. (2012). Scales of perception: public awareness of regional and neighborhood climates. Climatic Change , 111 (3 4), Saarinen, J., & Tervo, K. (2006). Perceptions and adaptation strategies of the tourism industry to climate change: the case of Finnish nature based tourism entrepreneurs. International Journal of Innovation and Sustainable Devel opment , 1 Satorra, A., & Bentler, P. M. (1988). Scaling corrections for chi square statistics in covariance structure analysis. 1988 Proceedings of the Business and Economic Statistics Section of the American Statistical Association Satorra, A., & Bentler, P. M. (1994). Corrections to test statistics and standard errors in covariance structure analysis. In A. von Eye & C.C. Clogg (Eds.), Latent Variable Analysis: Applications to Developmental Research e. Savaresi, A. (2016). The Paris Agreement: a new beginning?. Journal of Energy & Natural Resources Law , 34 Scheraga, J. D., & Grambsch, A. E. (1998). Risks, opportunities, and adaptation to climate change. Climate research , 11 Schmidt, A., Striegnitz, M., & Kuhn, K. (2014). Integrating regional perceptions into climate change adaptation: a transdisciplinary case study Regional environmental change , 14 Schultz, P. W. (1999). Changing behavior with normative feedback interventions: A field experiment on curbside recycling. Basic and applied social psychology , 21 Schumacker, R. E., & Lomax, R. G. (2004). s guide to structural equation modeling (2nd ed.). Mahwah, NJ: Lawrence Erlbaum Associates Inc. Scott, W. (1955). Reliability of content analysis: The case of nominal scale coding. Public Opinion Quarterly, 19

PAGE 219

205 Semenza, J. C., Hall, D. E., Wilson, D. J., Bontempo, B. D., Sailor, D. J., & George, L. A. (2008). Public perception of climate change: voluntary mitigation and barriers to behavior change. American journal of preventive medicine , 35 Seo, K. Y., Kim, W. H., Kim, H. A., & Lee, J. H. (2013). A Study Recognition of Science and Technology, Enviro nment, Climate Change in Korea. Journal of Climate Change Research , 4 Shear, M. D. (2017, June 1). Trump Will Withdraw U.S. From Paris Climate Agreement. The New York Times . Retrieved from https://www .nytimes.com/2017/06/01/climate/ trumpparis climate agreement.html. Sheffield, J., & Wood, E. F. (2012). Drought: past problems and future scenarios . New York, NY: Earthscan. Shi, J., Visschers, V. H., Siegrist, M., & Arvai, J. (2016). Knowledge as a driver of public perceptions about climate change reassessed. Nature Climate Change , 6 Shove, E. (2010). Beyond the ABC: climate change policy and theories of social change. Environment and planning A , 42 Sia, A. P., Hungerford, H. R., & Tomera, A. N. (1986). Selected predicto rs of responsible environmental behavior: An analysis. The Journal of Environmental Education , 17 (2), and Risk Analysis , 26 Simon, H. A. (1967). The business school a problem in organizational design. Journal of Management Studies , 4 Simpson, D. W. (1986). Triggered earthquakes. Annual Review of Earth and Planetary Sciences , 14 Singleton, R., & Straits, B. C. (2017). Approaches to Social Research (6th ed.). New York, NY: Oxford University Press. Sjöberg, L., Moen, B. E., & Rundmo, T. (2004). Explaining risk perception. An evaluation of the psychometric paradigm in risk perception research , 10 Slovic, P. (2000). Perception of Risk. In P. Slovic (E d.) , The Perception of Risk Sterling, VA: Earthscan Publications. Slovic, P., Finucane, M. L., Peters, E., & MacGregor, D. G. (2004). Risk as analysis and risk as feelings: Some thoughts about affect, reason, risk, and rationality. Risk analysis , 24 (2),

PAGE 220

206 Smit, B., Burton, I., Klein, R. J., & Wandel, J. (2000). An anatomy of adaptation to climate change and variabili ty. Climatic Change , 45 Smith, N., & Leiserowitz, A. (2012). The rise of global warming skepticism: Exploring affective image associations in the United States over time. Risk Analysis , 32 Song, I. G. (2015, October 8 ). Limited water supply in the South Chungcheong Province. The Hankyoreh . Retrieved from http://www.hani.co.kr/a rti/society/area/712051.html. (i n Korean). Spence, A., Poortinga, W., Butler, C., & Pidgeon, N. F. (2011). Perceptions of climate change and willingness to save energy related to flood experience. Nature climate change , 1 (1), Statistics Korea. (2017a). 2016 social indicators in South Korea . Retrieved from http://kostat.go.kr/portal/korea/kor_nw/2/6/5/index.board?bmode=read&aSeq=359629 . ( in Korean). Statistics Korea. (2017b). 2016 population in South Korea . Retrieved from http://kosis.kr/statisticsList/statisticsList_01List.jsp?vwcd=MT_ZTITLE&parentId=A . (in Korean). Statistics Korea. (2019). 2019 population in South Korea . Retrieved from http://kosis.kr/statHtml/statHtml.do?orgId=101&tblId=DT_1B040A3 . (in Korean) . Steiger, J. H., & Lind, J. C. (1980, May). Statistically based tests for the number of common factors. Paper presented at the annual meeting of the Psychometric Society, Iowa City, IA. Stern, N. (2006). What is the economics of climate change?. World Economics , 7 Stern, P. C. (2000). New environmental theories: toward a coherent theory of environmentally significant behavior. Journal of social issues , 56 Stevenson, K. T., Peterson, M. N., Bondell, H. D., Moore, S. E., & Carrier, S. J. (2014). Overcoming skepticism with education: interacting influences of worldview and climate change knowledge on perceived climate change risk among adolescents. Climatic ch ange , 126 Swim, J., Clayton, S., Doherty, T., Gifford, R., Howard, G., Reser, J., Stern, J., & Weber, E. (2009). Psychology and global climate change: Addressing a multi faceted phenomenon and set of challenges. A report by the American on the interface between psychology and global climate change . Washington, DC: American Psychological Association . Swim, J. K., Stern, P. C., Doherty, T. J., Clayton, S., Reser, J. P., Weber, E. U., ... & H oward, G. S s contributions to understanding and addressing global climate change. American psychologist , 66

PAGE 221

207 Takao, K., Motoyoshi, T., Sato, T., Fukuzondo, T., Seo, K., & Ikeda, S. (2004). Factors or floods in modern megalopolises: the case of the Tokai flood disaster in Japan. Journal of Risk Research , 7 (7 Tashakkori, A., & Creswell, J. W. (2007). Exploring the nature of research questions in mixed methods research. Journal of Mixed Methods Research , 1 Team, C. W., Pachauri, R. K., & Reisinger, A. (2007). Contribution of working groups I, II and III to the fourth assessment report of the intergovernmental panel on climate change. Geneva, Switzerland: IPCC. Terry, D. J., & Hogg, M. A. (2001). Attitudes, behavior, and social context: The role of norms and group membership in social influence processes. In J. P. Forgas & K. D. Williams (Eds.), Social influence: Direct and indirect processes , PA : Psychology Press. Thaker, J., Zhao, X., & Leiserowitz, A. (2017). Media Use and Public Perceptions of Global Warming in India. Environmental Communication , 11 Thieken, A. H., Kreibich, H., Müller, M., & Merz, B. (2007). Coping with floods: pre paredness, response and recovery of flood affected residents in Germany in 2002. Hydrological Sciences Journal , 52 Thurow, T. L., & Taylor Jr, C. A. (1999). The role of drought in range management. Journal of Range Management , 52 . Tinsley, H. E., & Weiss, D. J. (1975). Interrater reliability and agr eement of subjective judgments. Journal of Counseling Psychology , 22 Tol, R. S. (2011). Regulating knowledge monopolies: the case of the IPCC. Climatic Change , 108 (4), Tompkins, E. L., & Adger, W. N. (2005). Defining response capacity to enhance climate change policy. Environmental Science & Policy , 8 Trevino, L. K., Weaver, G. R., Gibson, D. G. , & Toffler, B. L. (1999). Managing ethics and legal co mpliance: what works and what hurts. California Management Review , 41 (2), 131 151. Tucker, L. R., & Lewis, C. (1973). A reliability coefficient for maxi mum likelihood factor analysis. Psychometrika , 38 Tunner Jr, J. F., Day, E., & Crask, M. R. (1989). Protection motivation theory: An extension of fear appeals theory in communication. Journal of Business Research , 19 Tvedt, T. (2010). Why England and not China and India? Water systems and the history of the Industrial Revolution. Jou rnal of Global History , 5 (

PAGE 222

208 Ullman , J. B., & Bentler , P. M. ( 2003 ) . Structural equation modeling . In J. A. Schinka & W. F. Velicer (Eds.), Handbook of psychology: Research methods in psychology ( Vol. 2 ., pp. John Wiley & Sons Inc . Ulrich, F. P. (1935). Man Made Earthquakes. Engineering News Record , 115 United Nations (UN). (1992). United Nations Framework Convention on Climate Change . Retrieved from ttps://unfccc.int/files/essential_background/background_publications _htmlpdf/application/pdf/conveng.pdf. United Nations Framework Convention on Climate Change (UNFCCC). (2017). Intended Nationally Determined Contributions (INDCs). Retrieved from http://u nfccc.int/ . United Nations Water. (2006). Water, A Shared Responsibility . The United Nations World Water Development Report 2 . Retrieved from http://www.unwater.org/publications/water shared responsibility/ . United States Geological Survey (USGS). (2017). Flood Definitions. Retrieved from https://ks.water.usgs.gov/flood definitions. Valls, J. F., & Sarda, R. (2009). Tourism expert perceptions for evaluating climate change impacts on the Euro Mediterranean tourism industry. Tourism Review , 64 Vance, A., Siponen, M., & Pahnila, S. (2012). Motivating IS security compliance: insights from habit and protection motivation theory. Information & Management , 49 Van de Ven, A. H. (2007). Engaged scholarship: A guide for organizational and social research . New York, NY: Oxford University Press. van der Linden, S. (2015). The social psychological determinants of c limate change risk perceptions: Towards a comprehensive model. Journal of Environmental Psychology , 41 , van der Linden , S. (2017). Determinants and Measurement of Climate Change Risk Perception , Worry , and Concern . In M.C. Nisbet, M. Schafer, E. Markowitz, S . Ho, S & J. Thaker (Eds.), The Oxford Encyclopedia of Climate Change Communication Oxford, UK: Oxford University Press. Available at http://dx.doi.org/10.2139/ssrn. 2953631. van der Pligt, J. (1996). Risk perception and self protective behavior. European Psychologist , 1 Vasileiadou, E., Heimeriks, G., & Petersen, A. C. (2011). Exploring the impact of the IPCC Assessment Reports on science. Environmental Science & Policy , 14 Vedung, E. (2017). Public policy and program evaluation . New York, NY: Routledge.

PAGE 223

209 Veronesi, M., Chawla, F., Maurer, M., & Lienert, J. (2014). Climate change and the willingness to pay to reduce ecological and health risks from wastewater flooding in urban centers and the environment. Ecological Economics , 98 Vining, J., & Ebreo, A. (1990). What makes a recycler? A comparison of recyclers and nonrecyclers. Environment and behavior , 22 Viscusi, W. K., & Zeckhauser, R. J. (2006). The perception and valuation of the risks of climate change: a rational and behavioral blend. Climatic change , 77 V örösmarty, C. J., Green, P., Salisbury, J., & Lammers, R. B. (2000). Global water resources: vulnerability from climate change and population growth. Science , 289 Wachinger, G., Renn, O., Begg, C., & Kuhlicke, C. (2013). The risk perception paradox implications for governance and communication of natural hazards. Risk analysis , 33 (6), Walther, G. R., Post, E., Convey, P., Menzel, A., Parmesan, C., Beebee, T. J., Fromentin, J., Hoegh Guldberg, O., & Bairlein, F. (2002). Ecological resp onses to recent climate change. Nature , 416 Wayne, G. P. (2013). Now available: a guide to the IPCC's new RCP emissions pathways. The Guardian. Retrieved from https://www.theguardian.com/environment/climate consensus 97 per cent/2013/aug/30/climate change rcp handy summary. Weber, E. U. (1997). Perception and Expectation of Climate Change: Precondition for Economic and Technological Adaptation. In M. H. Bazerman, D. M. Messick, A. E. Tenbrunsel & K. A. Wade Benzoni (Eds.), Environment, Ethics, and Behavior (pp. . San Francisco, CA: Jossey Bass. Weber, E. U. (2010). What shape s perceptions of climate change?. W iley Interdisciplinary Reviews: Climate Change , 1 Weimer, D., & Vining, A. (2011). Policy analysis: Concepts and practice (5th ed.). New York, NY: Routledge . Weinstein, N. D. (1989). Effects of personal experience on self protective behavior. Psychological bulletin , 105 West, S. G., Finch, J. F., & Curran, P. J. (1995). Structural equation models with nonnormal variables: Problems and remedies. In R. H. Hoyle (Ed.), Structural equation modeling: Concepts, issues, and applications Publications Inc. Whitmarsh, L. (2008). Are flood victims more concerned about climate change than other people? The role of direct e xperience in risk perception and behavioural response. Journal of risk research , 11

PAGE 224

210 Whitmarsh, L. (2009a). Behavioural responses to climate change: Asymmetry of intentions and impacts. Journal of environmental psychology , 29 Whitmar sh, L. (2009b). What's in a name? Commonalities and diff erences in public understanding Public understanding of science , 18 Wiegman, O., & Gutteling, J. M. (1995). Risk appraisal and ris k communication : Some empirical data from the Netherlands reviewed. Basic and Applied Social Psychology , 16 (1 Wilbanks, T. J., Kane, S. M., Leiby, P. N., Perlack, R. D., Settle, C., Shogren, J. F., & Smith, J. B. (2003). Possible responses to global climate change: integra ting mitigation and adaptation. Environment: Science and policy for sustainable development , 45 Wildavsky, A., & Dake, K. (1990). Theories of risk perception: Who fears what and why?. Daedalus , 119 Wilhite, D. A. (2000) . Drought as a natural hazard: Concepts and definitions. In D. A. Wilhite (Ed.), Droughts: A Global Assessment (pp. 3 18). London, UK : Routledge. Wilhite, D. A., & Glantz, M. H. (1985). Understanding: the drought phenomenon: the role of definitions. Water international , 10 Williams, R. (2016). Understanding and interpreting generalized ordered logit models. The Journal of Mathematical Sociology , 40 Winder, M., & Schindler, D. E. (2004). Climate change uncouples trophic interactions in an aquatic ecosystem. Ecology , 85 Wong, L. P. (2008). Data Analysis in Qualitative Research : a Brief Guide to Using Nvivo. Malaysian Family Physician, 3 (1), . World Bank. (2018). GDP per capita (constant 2010 US$) . Retrieved from https://data.worldbank.org/ indicator/NY.GDP.PCAP.KD?view=chart . World Commission on Environment and Development. (1987). Our Common Future . Oxford, UK: Oxford University Press. World Meteorological Organization. (2016). Guidelines on the definition and monitoring of extreme weather and climate events . Retrieved from https://www.wmo.int/pages/prog/ wcp/ccl/opace/opace2/documents/DraftversionoftheGuidelinesontheDefinitionandMonito ringofExtremeWeatherandClimateEvents.pdf. Worrell, E., Price, L., Martin, N., Hendriks, C., & Meida, L. O. ( 2001). Carbon dioxide emissions from the global cement industry. Annual review of energy and the environment , 26

PAGE 225

211 Xiao, C., & McCright, A. M. (2015). Gender differences in environmental concern: Revisiting the institutional trust hypothesis in the USA. Environment and Behavior , 47 Yates, J. F., & Stone, E. R. (1992). The risk construct. In J. F. Yates (Ed.), Wiley series in human performance and cognition. Risk taking behavior and: John Wiley. Yoo, S., & Lee, S. M. (1987). Management style and practice of Korean chaebols. California Management Review , 29 Young, O. (1991). Political Leadership and Regime Formation: On the Development of Institutions in International Society. International Organization , 45 (3), 281 309. Yu, H., Wang, B., Zhang, Y. J., Wang, S., & Wei, Y. M. (2013). Public perception of climat e change in China: results from the questionnaire survey. Natural hazards , 69 Yung, Y. F., & Bentler, P. M. (1996). Bootstrapping techniques in analysis of mean and covariance structures. In G. A. Marcoulides & R. E. Schumacker (Eds.), Advanced structural equation modeling: Issues and techniques (pp. 195 226). Mahwah, NJ: Lawrence Erlbaum Associates Inc. Zaalberg, R., Midden, C., Meijnders, A., & McCalley, T. (2009). Prevention, adaptation, and threat denial: Flooding experiences in the Netherlands. Risk Analysis , 29 Zahran, S., Brody, S. D., Grover, H., & Vedlitz, A. (2006). Climate change vulnerability and policy support. Society and Natural Resources , 19 Zaleskiewicz, T., Piskorz, Z., & Borkowska, A. (2002). Fear or money? Decisions on insuring oneself against flood. Risk, Decision and Policy , 7 Zamani, G. H., Gorgievski Duijvesteijn, M. J., & Zarafshani, K. (2006). Coping with drought: towards a multilevel understanding based on conservat ion of resources theory. Human Ecology , 34 Zillman, J. W. (2007). Some observations on the IPCC assessment process 1988 2007. Energy & Environment , 18

PAGE 226

212 APPENDIX A Conceptual map for v ariables Concepts Definitions Indicators Related studies Weather The short term conditions (changes) of the atmosphere in terms of temperature, precipitation, wind, etc. Auxiliary concept NASA (2011); Lombardi and Sinatra (2012); Rebetez (1996); Weber (2010); Whitmarsh (2008) Climate Average weather conditions over a long period of time ( a statistical phenomenon) Climate change A change in the typical or average weather caused by (direct and indirect) human activities or natural origin Auxiliary concept IPCC (2001); NASA (2011); Team, Pachauri, and Reisinger (2007); UN (1992); Whitmarsh (2009b) Personal climate change knowledge Awareness of climate change itself, the understanding of the causes and consequences of climate change, and the familiarity with coping strategies acquired by a person through experience or education (1) Awareness of climate change (2) The understanding of the cause of climate change (3) The familiarity with a government GHG reduction plan Banks (1993); Halady and Rao (2010); onnor, Bard, and Fisher (1999); Shi et al . (2016) Personal climate change experience Experiences of the long term features of climate change (e.g., variations in temperature, precipitation, etc.) or short term extreme weather events (e.g., drought and flood) intensified by climate change (1) Experienc es of variations in temperature, precipitation, etc. over a long period of time (2) Experience of drought (3) Experience of flood Jentsch, Kre yling, and Beierkuhnlein (2007); NASA (2011); Meehl et al . (2000); Mirza (2003); Revi (2008); Rosenzweig et al . (2001)

PAGE 227

213 Concepts Definitions Indicators Related studies Personal climate policy assessment A subjective evaluation of government policies addressing climate change in terms of necessity, validity, efficiency, effectiveness, etc. (1) Assessment of drought policy (2) Assessment of flood policy Crabbé and Leroy (2012); Vedung (2017) Climate change risk The probability of exposure to perilous consequences on the health or living conditions of human s and ecosystems brought by anthropogenic or natural origin c limate change (or conditions produced by previous climate change) Auxiliary concept Böhm and Pfister (2001); Jones (2001); Pachauri et al . (2014); UN (1992) Climate change risk perception A subjective evaluation of the probability of clim ate change (or conditions produced by previous climate change) happening and the severity of and vulnerability to its adverse consequences (1) The seriousness of climate change impacts Leiserowitz (2006); Rogers (1983); Sjöberg, Moen, and Rundmo (2004) Personal behavioral responses Actual actions to reduce GHG emissions (e.g., saving energy), to adapt to climate change, or to support climate policies and environmental organizations (e.g., voting and donation) (1) The number of behavioral responses to address climate change (e.g., saving energy, reducing the use of disposable items, and using renewable energy) De ng, Wang, and Yousefpour (2017); Grothmann and Patt (2005); Masud et al . (2016); Whitmarsh (2009b)

PAGE 228

214 APPENDIX B Key questions of the 2016 s urvey in South Korea V a riables Question s Experience of drought or flood Question 2 3. Have you experienced a ny property damages caused by flood or drought in the past? 1) No experience 2) By drought 3) By flood 4) By both drought and flood Experience of climate change Based on your experiences in the last ten years, Question 4. Average temperature is _______________ rising. Question 5. Precipitation is _______________ changing. Question 6. The danger of flood is _______________ increasing. Question 7. The danger of drou ght is _______________ increasing . 1) Definitely 2) Somewhat 3) Little 4) Not Awareness of climate change Question 1. Do you know about global warming and its impacts? 1) Yes 2) No T he cause of climate change Question 3 . What do you think of the contribution of human activities to climate change? [ ] Please provide numeric scores from 0% up to 100%. The understanding of a GHG reduction plan Question 9 . To what extent do you know about the Korean uction plan of Green House Gas (GHG) emissions? 1) Precisely 2) Moderately 3) Hardly 4) Not Drought policy assessment Question 17 . What do you think of drought policies for addressing climate change? 1) Very poor 2) Poor 3) Fair 4) Good 5) Excellent Flood policy assessment Question 16 . What do you think of flood policies for addressing climate change? Climate change risk perception Question 2 . What do you think of the adverse impacts of climate change brought by global warming? 1) Strongly serious 2) Moderately s erious 3) Hardly serious 4) Not serious

PAGE 229

215 V a riables Question s Behavioral responses to alleviate climate change Question 2 0 . What kind of behavi oral efforts are you currently making to address climate change? Please select all items you have been practicing. 1) Saving energy 2) Saving water 3) Using public transportation instead of a privately owned car 4) Reducing the use of disposable items 5) Utilizing renewable energy 6) Enjoying a vegetarian diet instead of a meat diet 7) Donations to environmental organizations 8) Voting based on environmental related election promises Gender Basic Inf . What is your gender? 1) Male 2) Female Age Basic Inf. What is your age? [ ] Please express in numbers. Education Question 24 . What is the highest level of education you have completed? 1) Under middle school 2) High school 3) Undergraduate 4) Graduate and above Income Question 25 . How much was your average monthly income before tax last year? 1) Inc < KRW 1million (US$ 909) 2) KRW 2million (US$ 1,818) 3) KRW 3million (US$ 2,727) 4) KRW 4million (US$ 3,636) 5) K KRW 5million (US$ 4,545) 6) KRW 6million (US$ 5,455) 7) KRW 6million (US$ 5,455) < Inc

PAGE 230

216 APPENDIX C R esults of SEM regression analyses wi th different estimation methods (or options) in the national level analysis

PAGE 231

217 APPENDIX D Direct, ind irect, and total effects of independent and control variables on behavioral responses to mitigate climate change Independent v ariables The number of behavioral responses Climate change risk perception Total .260*** Direct .260*** Drought experience Total .313* Direct .300 Indirect .013 Flood experience Total .118 Direct .118 Indirect .000 Climate change experience Total .028*** Indirect .028*** Awareness of climate change Total .056* Indirect .056* Cause of climate change Total .001** Indirect .001** Familiarity with a GHG reduction plan Total .206*** Direct .206*** Assessment of drought policy Total .017* Indirect .017* Assessment of flood policy Total .005 Indirect .005 Gender (female) Total .112 Direct .078 Indirect .034** Age Total .018*** Direct .017*** Indirect .001 Education Total .029 Direct .027 Indirect .002 Income Total .045 Direct .044 Indirect .001 Area Total .037 Direct .036 Indirect .001 *p<.05, ** p<.01, ***p<.001 (two tailed test)

PAGE 232

218 APPENDIX E Relationship between p ersonal factors and socio demographic characteristics in the national and local level analyses E_1. National level analysis Independent Variables Dependent variables Experience Knowledge Policy assessment Drought experience Flood experience Climate change experience Awareness of climate change Cause of climate change Familiarity with a GHG policy Assessment of drought policy Assessment of flood policy Coeff. Coeff. Coeff. Coeff. Coeff. Coeff. Coeff. Coeff. Gender .488** (.166) (.614**) .491*** (.129) (.612***) .315** (.094) .135 (.254) (.874) .928 (.941) .511*** (.099) (.600***) .005 (.098) (1.004) .294** (.101) (.745**) Age .007 (.006) (1.007) .019*** (.005) (1.019***) .014*** (.004) .001 (.010) (.999) .098** (.035) .023*** (.004) (1.023***) .021*** (.004) (1.021***) .015*** (.004) (1.015***) Education .092 (.136) (.912) .152 (.101) (.859) .131 (.078) .317 (.212) (1.373) 2.825*** (.740) .204* (.081) (1.227*) .093 (.084) (.911) .187* (.082) (.829*) Income .058 (.049) (.944) .003 (.039) (1.003) .028 (.030) .221* (.087) (1.247*) .072 (.291) .050 (.031) (1.051) .051 (.030) (.950) .032 (.031) (.968) Area .276* (.135) (1.318*) .040 (.105) (1.041) .083 (.087) .046 (.227) (.955) .951 (.811) .094 (.080) (1.098) .062 (.082) (1.063) .073 (.083) (1.076) Constant 1.781 1.499 8.963 1.427 73.186 (Pseudo) R 2 .016 .021 .020 .027 .020 .025 .012 .010 N 1,500 1,500 1,484 1,500 1,500 1,500 1,500 1,500 Model Logit Logit OLS Logit OLS Ordered logit Ordered logit Ordered logit Note 1. All regression analyses are with robust standard errors. 2. The first and second parentheses display robust standard errors and odds ratios, respectively. 3. Multicollinearity, influential outliers, specification errors, and other significant statistical problems are not detected. *p<.05, **p<.01, ***p<.001 (two tailed test)

PAGE 233

219 E_2. Local le vel analysis Independent Variables Dependent variables Experience Knowledge Policy assessment Water problem experience Awareness of water quantity condition Exposure to water saving information Familiarity with a subsidy policy Assessment of water supply policy Assessment of water infra. policy Coeff. Coeff. Coeff. Coeff. Coeff. Coeff. Gender .098 (.307) (1.103) .323 (.231) (1.381) .164 (.220) (1.178) .202 (.192) (1.224) .340 (.237) (1.404) .185 (.209) (.831) Age .023 (.015) (.978) .002 (.012) (1.002) .003 (.011) (1.003) .020* (.009) (1.020*) .001 (.011) (.999) .018 (.010) (1.019) Education .019 (.314) (.981) .261 (.206) (1.298) .088 (.206) (1.092) .516** (.169) (1.675**) .299 (.229) (.741) .076 (.195) (.926) Income .166 (.115) (.847) .185* (.090) (.832*) .043 (.086) (1.044) .091 (.077) (1.095) .160 (.100) (1.174) .082 (.087) (1.086) Area .699* (.274) (2.013*) .873*** (.224) (2.393***) 1.677*** (.198) (5.350***) .052 (.171) (1.053) .259 (.231) (1.296) 1.306*** (.193) (3.692***) Constant .630 1.127 (Pseudo) R 2 .027 .035 .114 .010 .012 .070 N 508 504 508 508 508 507 Model Logit Ordered logit Logit Ordered logit Ordered logit Ordered logit Note 1. All regression analyses are with robust standard errors. 2. The first and second parentheses display robust standard errors and odds ratios, respectively. 3. Multicollinearity, influential outliers, specification errors, and other significant statistical problems are not detected. *p<.05, **p<.01, ***p<.001 (two tailed test)

PAGE 234

220 APPENDIX F Key questions of the 2015 s urvey in north west area of South Ch ungcheong Province V a riables Question s Experience of water supply problems Question 19. Have you experienced any water supply problems (e.g., limited water supply, low pressure, etc.) during this drought period? 1) Yes 2) No The condition of water quantity at the national level Question 1. What do you think of water quantity in South Korea? 1) Very abundant 2) Abundant 3) Scarce 4) Very scarce Information about water saving Question 12. Have you been exposed to a campaign, publicity, or education related to water saving? 1) Yes 2) No The understanding of a subsidy policy related to water saving Question 13 . To what extent do you know about the Korean to water saving during a drought period? 1) Precisely 2) Moderately 3) Hardly 4) Not Water supply assessment Question 4 . What do you think of in convenience in using tap water in your area? 1) Very serious 2) Moderately serious 3) Hardly serious 4) Not serious Water infrastructure assessment Question 5. What do you think of water supply infrastructure (e.g., tap water facilities) in your area? 1) Very good 2) Good 3) Poor 4) Very poor Drought risk perception Question 3. What do you think of the seriousness of drought in your area? 1) Very serious 2) Moderately s erious 3) Hardly serious 4) Not serious Behavioral responses to save water (Daily actions) Question 14. Regarding tap water savings during this drought period compared to non drought period, making an effort to use a glass with water when brushing my teeth. making an effort to use a wash basin with water when washing my face. 1) Strongly 2) Moderately 3) Hardly 4) Not

PAGE 235

221 making an effort to shorten shower time. Question 14_4 making an effort to turn off the Quest ion 14_5 making an effort to do laund ry in a larger quantity. Question 14_6 making an effort to use a washing up bowl when washing dishes. Question 14_7 making an effort to inspect a leakage of water. Gender Basic Question 4 . What is your gender? 1) Male 2) Female Age Basic Question 5 . What is your age? [ ] Please express in numbers. Education Basic Question 6 . What is the highest level of education you have completed? 1) Under middle school 2) High school 3) Undergraduate and above Income Basic Question 8 . How much was your average monthly income before tax last year? 1) Inc < KRW 1million (US$ 909) 2) KRW 2million (US$ 1,818) 3) KRW 3million (US$ 2,727) 4) KRW 4million (US$ 3,636) 5) KRW 5million (US$ 4,545) 6) KRW 6million (US$ 5,455) 7) KRW 6million (US$ 5,455) < Inc Region Basic Question 1 . Where do you live in? 1) Dangjin city 2) Boryeong city 3) Seosan city 4) Seocheon county 5) Yesan county 6) Cheongyang county 7) Taean county 8) Hong seong county

PAGE 236

222 APPENDIX G R esults of SEM regression analyses with different estimation methods (or options) in the local level analysis G_1. Daily action (Range from 0 to 21) Dependent variables Independent variable SEM/ML SEM/ML/ ROBUST (SBENTLER) SEM/ML/ BOOTSTRAP SEM/ADF Coeff. Std. Err. Coeff. Std. Err. Coeff. Std. Err. Coeff. Std. Err. Drought risk perception Water problem experience .094 .079 .094 .076 .094 .083 .061 .064 Awareness of water quantity condition .206*** .059 .206*** .058 .206*** .059 .226*** .050 Exposure to water saving information .198** .057 .198*** .056 .198*** .049 .202*** .052 Assessment of water supply policy .007 .063 .007 .059 .007 .062 .016 .049 Assessment of water infrastructure policy .186*** .049 .186*** .051 .186*** .048 .192*** .044 Gender (female) .124* .058 .124* .058 .124* .053 .111* .052 Age .001 .003 .001 .003 .001 .002 .001 .002 Education .022 .055 .022 .056 .022 .056 .040 .051 Income .047* .024 .047* .022 .047* .023 .048* .019 Area .045 .060 .045 .058 .045 .055 .039 .053 Constant 2.787 2.787 2.787 2.872 Behavioral responses Drought risk perception .040 .243 .040 .250 .040 .282 .032 .224 Self vulnerability perception .101* .049 .101 .051 .101* .047 .101* .043 Water problem experience .407 .435 .407 .484 .407 .500 .384 .379 Exposure to water saving information 1.049** .330 1.049** .372 1.049** .336 1.065** .317 Familiarity with a subsidy policy .398** .151 .398* .161 .398** .138 .413** .139 Gender (female) .568 .325 .568 .339 .568 .341 .552 .294 Age .070*** .016 .070*** .015 .070*** .014 .064*** .013 Education .301 .311 .301 .324 .301 .320 .298 .275 Income .206 .136 .206 .115 .206 .120 .172 .100 Area .931** .320 .931** .337 .931** .294 1.000*** .283 Constant 11.834 11.834 11.834 12.277 n 501 501 501 501 Model fits Acceptable thresholds p>.05 ( p<.05 (11.488, 5) p>05 (10.493, 5) p< .05 (11.488, 5) p<.05 (13.335, 5) RMSEA <.07 (or .06) 0.051 0.047 0.051 0.058 SRMR <.08 0.011 0.011 0.011 0.016 CFI >.95 0.951 0.959 0.951 0.939 TLI >.95 0.755 0.793 0.755 0.693 *p<.05, **p<.01, ***p<.001 (two tailed test)

PAGE 237

223 G_2. Device utilization (Range from 0 to 3)

PAGE 238

224 APPENDIX H Direct, ind irect, and total effects of independent and control variables on behavioral responses to adapt to drought Independent v ariables Daily action model Device utilization model Drought risk perception Total .040 .053 Direct .040 .053 Self vulnerability perception related to limited water supply Total .101 .078*** Direct .101 .078*** Water problem experience Total .403 .097 Direct .407 .092 Indirect .004 .005 Awareness of water quantity condition Total .008 .011 Indirect .008 .011 Ex posure to water saving information Total 1.057** .003 Direct 1.049** .013 Indirect .008 .010 Familiarity with a subsidy policy related to drought Total .398* .119** Direct .398* .119** Assessment of water supply policy Total .000 .000 Indirect .000 .000 Assessment of water infrastructure policy Total .007 .010 Indirect .007 .010 Gender (female) Total .573 .044 Direct .568 .037 Indirect .005 .007 Age Total .070*** .001 Direct .070*** .001 Indirect .000 .000 Education Total .302 .267** Direct .301 .266** Indirect .001 .001 Income Total .204 .138*** Direct .206 .141*** Indirect .002 .003 Area Total .929** .549*** Direct .931** .551*** Indirect .002 .002 *p<.05, ** p<.01, ***p<.001 (two tailed test)

PAGE 239

225 APPENDIX I Main questions of the c limate c hange expert interview 1) Can you please provide your full name, academic background, professional title, and research center that you work for? 1) How would you describe the mission/purpose of your research center and your job at the center? In this interview, individuals refer to the general public. 2) How would you describe the importance of the role of individuals in addressing climate change? 3) How would you descr as flood, drought, and heat wave, on their climate change risk perception? 4) term climate change experiences, such as the rising of average temperature and the change in precipitation in the last ten years, on their climate change risk perception? 5) How would you describe the impacts of their climate change risk perception? 6) How wou (e.g., drought and flood policies) on their climate change risk perception? 7) beha vioral responses to address climate change? 8) behavioral responses to address climate change? 9) especially at your research center?

PAGE 240

226 APPENDIX J Part of the c odebook for interview transcription Coding I nstructions: Coders should before coding . When coding, Codebook: INDROLEIM climate change o IRNIM Mentions that their role is not at all important in addressing climate change o IRSWI Mentions that their role is somewhat impo rtance in addressing climate change o IRVRI Mentions that their role is very important in addressing climate change INDROLECM o IRLIOSR roles, such as governments, industries, and environmental organizations, in addressing climate change o IRAIAOS NG Mentions that their role is as important as other stakeho roles (not including the o IRAIAOSR o IRMIOSNG Mentions that their r ole is more important than other stakeho roles (not including the o IRAIAGR Mentions that th eir role is as important as the addressing climate change o IRMIOSR Mentions that their role is more important than other sta roles, including the INDROLE PB role on the burden of policymakers in designing and implementing climate policies o IRDP B Mentions that their role decreases the burden of policymakers in designing and implementing climate policies

PAGE 241

227 o IRNPB Mentions that their role has nothing to do with the burden of policymakers in designing and implementing climate policies o IRIPB Menti ons that their role increases the burden of policymakers in designing and implementing climate policies STEXPERIENCE experiences on their climate change risk perception o SENECP Mentions of i ndividuals believing climate change is less severe when they experience extreme weather events o SEPOCP Mentions of individuals believing climate change is more severe when they experience extreme weather events o SEORCP Mentions of other relationships bet ween climate change risk perception and extreme weather event experiences LTEXPERIENCE term climate change experiences on their climate change risk perception o LENECP Mentions of individuals believing climat e change is less severe when they experience long term climate change (e.g., the rising of average temperature and the change in precipitation) o LEPOCP Mentions of individuals believing climate change is more severe when they experience long term climate change (e.g., the rising of average temperature and the change in precipitation) o LEORCP Mentions of other relationships between climate change risk perception and long term climate change experiences KNOWLEDGE Mentions of the impacts of nowledge related to climate change on their climate change risk perception o KNNECP decreases their climate change risk perception o KNPOCP te change increases their climate change risk perception o KNORCP Mentions of other relationships between climate change risk perception and knowledge related to climate change ASSESSMENT Mentions of climate policy on their climate change risk perception o PANECP Mentions that individuals who perceive higher climate change risks will not support climate policy (or individuals who perceive lower climate change risks will support climate policy)

PAGE 242

228 o PAPOCP Mentions that individuals who perceive higher climate change risks will support climate policy (or individuals who perceive lower climate change risks will not support climate policy) o PAORCP Mentions of other relationships between climate change risk pe rception and climate change policy support BERESPONSE Mentions of perception on their behavioral responses to address climate change o CPNEBR Mentions that individuals who perceive higher climate change ri sks are less likely to take action to address climate change (or individuals who perceive lower climate change risks are more likely to take action to address climate change) o CPPOBR Mentions that individuals who perceive higher climate change risks are m ore likely to take action to address climate change (or individuals who perceive lower climate change risks are less likely to take action to address climate change) o CPORBR Mentions of other relationships between climate change risk perception and behavi oral responses to address climate change ALTERNATIVE Mentions of policy, programs, or education that contribute to o FIENBR Mentions of financial incentives (e.g., tax reduction linke d to less driving, subsidy and free parking service for electric cars, and electricity or water bill reduction based on an annual energy or water usage pattern) o FPENBR Mentions of financial penalties (e.g., higher taxes on gas guzzlers, temporary water p rice hike during drought periods, and fines for households which use more water than usual during drought periods) o ISENBR Mentions of information systems (e.g., climate change specific web based homepages or digital library systems, off line libraries, a nd call centers) o PCENBR Mentions of the enhancement of propaganda, public campaign, etc. o EDENBR Mentions of education included in the regular curriculums (e.g., primary, secondary, or tertiary education) o CPENBR Mentions of civil participatory decisio n making processes (e.g., public participation in designing climate policy and the above mentioned incentive or penalty policy) o MCENBR Mentions of the promotion of multidimensional communication (e.g., seminars, workshops, and public hearings)